IoT-Cloud Service Optimization in Next Generation Smart Environments

Size: px
Start display at page:

Download "IoT-Cloud Service Optimization in Next Generation Smart Environments"

Transcription

1 1 IoT-Clod Service Optimization in Next Generation Smart Environments Marc Barcelo, Alejandro Correa, Jaime Llorca, Antonia M. Tlino, Jose Lopez Vicario, Antoni Morell Universidad Atonoma de Barcelona, Spain, {marc.barcelo, alejandro.correa, jose.vicario, Nokia Bell Labs, NJ, USA, {jaime.llorca, Universita degli Stdi di Napoli Federico II, Italy, Abstract The impact of the Internet of Things (IoT) on the evoltion towards next generation smart environments (e.g., smart homes, bildings, cities) will largely depend on the efficient integration of IoT and clod compting technologies. With the predicted explosion in the nmber of connected devices and IoT services, crrent centralized clod architectres, which tend to consolidate compting and storage resorces into a few large data centers, will inevitably lead to excessive network load, end-toend service latencies, and overall power consmption. Thanks to recent advances in network virtalization and programmability, highly distribted clod networking architectres are a promising soltion to efficiently host, manage, and optimize next generation IoT services in smart environments. In this paper, we mathematically formlate the service distribtion problem (SDP) in IoT- Clod networks, referred to as the IoT-CSDP, as a minimm cost mixed-cast flow problem that can be efficiently solved via linear programming. We focs on energy consmption as the major driver of today s network and clod operational costs and characterize the heterogeneos set of IoT-Clod network resorces according to their associated sensing, compting, and transport capacity and energy efficiency. Or reslts show that, when properly optimized, the flexibility of IoT-Clod networks can be efficiently exploited to deliver a wide range of IoT services in the context of next generation smart environments, while significantly redcing overall power consmption. Index Terms Internet of Things, clod, edge compting, service optimization, energy efficiency, smart cities I. INTRODUCTION The Internet of Things (IoT) refers to the network of objects, devices, machines, vehicles, bildings, and other physical systems with embedded sensing, compting, and commnication capabilities, that sense and share real-time information abot the physical world [1], [2]. When connecting the IoT to the Clod, vast amonts of data collected from mltiple locations can be processed and analyzed to create meaningfl information for the end sers. At the same time, the intrinsic limitations of lightweight mobile devices (e.g., battery life, processing power, storage capacity) can be alleviated by taking advantage of the extensive resorces in the Clod [3], [4]. The reslting IoT-Clod paradigm enables a new breed of services and applications (e.g., health system monitoring, traffic control, energy management, vehiclar networking), expected to define the essence of next generation smart environments (e.g., smart homes, smart bildings, smart cities, smart grids) [5], [6]. However, with the predicted explosion in the nmber of IoT services and connected devices, traditional centralized clod architectres, in which compting and storage resorces are concentrated in a few large data centers, will inevitably lead to excessive network load, end-to-end service latencies, and nbearable energy costs [7], [8]. In order to meet the tight QoS reqirements associated with real-time IoT applications while maximizing overall efficiency, clod architectres are becoming increasingly distribted [7], with the presence of small clod nodes at the edge of the network, referred to as clodlets [9], micro-clods [10], fog nodes [11], or simply edge clod nodes [12]. The reslting clod infrastrctre may be organized into hierarchical layers, each with different compting and storage capabilities [13]. In addition, thanks to recent advances in mobile compting [14] and device layer virtalization [15], even end devices can become part of this highly distribted networked compte platform, creating what we refer to as IoT-Clod networks. These are highly virtalized heterogeneos platforms that offer compting, storage, and networking services with nprecedented distribtion and proximity to the end ser. Compared to traditional centralized clods, IoT-Clod networks provide increased flexibility in the allocation of resorces to IoT services, and a clear advantage in meeting their stringent latency, mobility, and locationawareness constraints. In the IoT-Clod network paradigm, services take the form of virtal network slices of the shared physical infrastrctre, which can flexibly and elastically consme compte, storage, and network resorces according to changing service reqirements. A key aspect driving both performance and efficiency is the actal placement of the service fnctions, as well as the roting of network flows throgh the appropriate fnction instances. In the context of clod networks, [16] introdced the clod service distribtion problem (CSDP), where the goal is to find the placement of virtal fnctions and the roting of network flows that meets QoS reqirements, satisfies resorce capacities and minimizes overall infrastrctre cost. The athors provide a network-flow based linear programming soltion that optimizes the distribtion of clod services with arbitrary fnction relationships (e.g., service chaining) over a distribted clod network. However, the work in [16] does not take into accont the increased flexibility that arises when introdcing the access network and the device layer into the virtalized infrastrctre, aspects that are critical for the efficient delivery of IoT services [17]. In this paper, we formalize the IoT-CSDP as a service placement and resorce allocation problem that goes beyond

2 2 traditional information services and clod architectres to inclde next generation IoT services and IoT-Clod infrastrctres. Or contribtions can be smmarized as follows: We introdce a flexible mathematical model for IoT- Clod networks that characterizes the capacity, efficiency, and reliability of sensing, compting, and transmission resorces across end device, access, and clod layers. Bilding on the clod service model introdced in [16], we characterize a generic IoT service via a directed rooted graph that encodes the relationship between the service fnctions that act on the sorce information flows to create the final agmented information that needs to be delivered to the end sers. We formally define the IoT service distribtion problem (IoT-CSDP) as the problem of finding the placement of IoT service fnctions and roting of network flows across the IoT-Clod infrastrctre that satisfy end ser demands while minimizing overall operational cost. We formlate the IoT-CSDP as a minimm cost mixed-cast network information flow problem on a properly agmented graph that allows captring nicast and mlticast flows, and admits optimal polynomial-time soltions. We evalate the soltion to the IoT-CSDP in an illstrative set of next generation smart environments that reqire the joint orchestration of device, access, and clod layers, and show the impact of the most relevant system parameters. Or reslts show that the proposed optimization model and associated flow-based soltion enables flly exploiting the flexibility of IoT-Clod networks to efficiently host, manage, and optimize a wide range of IoT services in the context of next generation smart environments. When compared to conventional centralized clod approaches, or soltion achieves p to 80% overall power consmption redctions while garanteeing mch more stringent end-to-end latency constraints. The remainder of this paper is organized as follows: Section II reviews related work. Section III presents the IoT- Clod network paradigm. Section IV describes the system model. Section V introdces the IoT-CSDP and its flowbased formlation. Section VI presents simlation reslts from the soltion to the IoT-CSDP in the context of illstrative smart environments: smart cities, smart bildings, and smart transportation systems. Finally, Section VII smmarizes the paper and presents the main conclsions. II. RELATED WORK In the context of integrating end devices into the Clod, special attention has been given to wireless sensors networks (WSNs) de to their relevance in the IoT [18]. In [19], the athors present a theoretical model for -Clod architectres and stdy the performance improvements that can be obtained over traditional WSN architectres. In [20], the athors describe a -Clod architectre that makes se of the Contiki Operating System to provide end sers with access to the data acqired by different heterogeneos sensing infrastrctres. In [21], the athors propose WSN-SOrA, a service oriented architectre that orchestrates service provisioning for embedded networked systems in large scale WSNs. In [22], the athors propose a sensory data processing framework that redces the amont of data forwarded to the Clod by processing data at the gateway layer sing monitoring, filtering, prediction, compression, and recommendation techniqes. In [23], the athors address sensor data seflness and WSN reliability by proposing a WSN-Clod integration scheme in which WSN gateways selectively transmit information to the Clod based on the time and priority featres of the sers reqested data, and wireless sensors execte a priority-based sleep schedling algorithm to redce their power consmption. In [24], a trst and reptation calclation and management system with athentication is proposed to address some of the secrity concerns that arise when integrating WSNs into the Clod. In [25], the athors propose a position-based sleepschedling mechanism that redces the active time of dtycycled sensors in order to enhance the lifetime of wireless sensors connected to the Clod. This problem is also addressed in [26], where the athors propose a management framework to control the dty cycle of sensors in IoT environments that may be performing mltiple tasks with different qality-ofinformation reqirements. In addition to wireless sensors, the efficient integration of smart devices, sch as smartphones and connected vehicles, has also attracted the attention of the research commnity. The athors in [27] analyze the se of smartphones as shared edge compting devices, taking advantage of their increasing processing capabilities. In [28], software defined networking (SDN) and edge compting technologies are combined to enable programmable vehiclar ad hoc networks (VANETs). The work in [29] presents a model for fog compting architectres with edge or fog nodes forming an intermediate layer between device and clod layers. The athors show how the fog layer enables relevant energy savings when spporting IoT applications, bt do not provide any optimization strategy to distribte the service fnctions over the entire infrastrctre. In [30], the athors develop a federation strategy among IoT devices and clods so they can share their compting resorces in order to maximize the total nmber of exected tasks. In [31], the athors present an IoT resorce management system in which micro data centers co-located with network gateways at the edge of the network perform billing, pricing, and resorce reservation for IoT services. They also propose a probabilistic resorce estimation model that takes into accont the flctating connectivity behavior of IoT devices [32]. While existent literatre provides a significant amont of stdies describing alternative models and architectres that illstrate the advantages, limitations, and challenges associated with IoT-Clod networks, to the best of or knowledge, this is the first work that mathematically formalizes the problem of optimal distribtion of generic IoT services over IoT-Clod networks, taking into accont the heterogeneos natre of sensing, transmission, and compting resorces across the physical infrastrctre, as well as the niqe fnction interrelationships, mixed-cast flow natre, and tight QoS reqirements of IoT services.

3 3 III. IOT-CLOUD NETWORKS IoT-Clod networks reslt from the convergence of distribted clod networks and the IoT. These can take advantage of the biqitos sensing capabilities of the IoT and the virtally nlimited compting resorces of the Clod to offer a new class of services that create agmented information from the clod-based analysis of IoT-based data. In IoT-Clod networks, sensors, smartphones, connected vehicles, are not simple endpoints, bt smart sensing, storing, and compting resorces that can be jointly orchestrated with the rest of the clod infrastrctre [33]. In this new paradigm, and thanks to recent advances in network fnctions virtalization (NFV) and software defined networking (SDN) technologies [34], service fnctionality can be dynamically allocated across the reslting highly distribted platform and flows can be roted throgh the appropriate service fnctions in order to maximize end devices battery life, optimize service performance, and minimize overall operational cost. The main advantages of IoT-Clod networks with respect to traditional centralized clod approaches are: Low latency: service fnctions can be placed at the edge of the network in close proximity to the end sers to spport real-time services. High reliability: service fnctions can be replicated across a highly distribted platform for increased falt tolerance and disaster recovery. Redced operational cost: service fnctions that reqire large inpts or prodce large otpts can be placed close to their respective sorces and/or destinations for redced network load and associated operational cost. High flexibility: virtalization technologies allows sharing the heterogeneos physical infrastrctre among mltiple services that can elastically tap into a rich pool of resorces withot the need of dedicated deployments. Location awareness and mobility spport: the crrent location of mobile sers can be leveraged to provide personalized contextal services as well as service mobility. Scalability: the distribtion of compting and storage resorces close to the sorces of information allows scaling the nmber of connected devices and services withot satrating the transport network. IoT-Clod networks hence emerge as an ideal platform for the implementation of IoT services in the context of a wide range of smart environments, sch as smart grids, smart mobility, smart bildings, and smart cities. For instance, in smart grids, the data collected by smart meters can be preanalyzed at the edge of the network redcing the amont of data sent to clod data centers, saving both compting and transport resorces, and significantly redcing service latency. In smart mobility services, the analysis of data collected at the edge of the network can significantly increase responsiveness to sdden events, sch as vehicle collisions, improving the efficiency and safety of transportation networks. In smart bildings, while centralized clod resorces can be sed for complex data analysis, edge clod nodes are ideally sited for low-latency real-time physical system control and actation. Health-Care, Critical Infrastrctres Monitoring Clod Network Pblic Transport, Polltion, Traffic Monitoring Smart Devices Connected Vehicles Fig. 1: IoT-Clod network reslting from the integration of IoT devices into the clod network infrastrctre. IV. SYSTEM MODEL After describing the sitability of IoT-Clod networks for the delivery of IoT services in ftre smart environments, we now introdce the mathematical network and service models that will be sed to optimize the distribtion of IoT services over IoT-Clod networks. A. Network model We refer to an IoT-Clod network as the converged platform that reslts from the integration of programmable IoT devices into the clod infrastrctre (Fig. 1). We model an IoT- Clod network as a directed graph G = (V, E) with V vertices and E edges representing the set of network nodes and links, respectively. A node in V may represent an end device, an access point, or a clod node (at possibly different hierarchical layers). Each node is characterized by its energy resorces (e.g., power grid, battery), processing resorces (e.g., processor, microprocessor), and data acqisition or sensing resorces (e.g., camera, sensors, I/O interfaces). We denote by c pr and c sn the data processing and sensing capacities (in bits per second or bps) at node 2V, and by e pr and e sn the data processing and sensing nit energy costs (in Watts per bps) at node 2V, respectively. Nodes are interconnected via wireless or wireline links, each characterized by their transmission capacity and nit energy cost. We se c tr v and e tr v to denote the capacity (in bps), and the nit energy cost (in Watts per bps) of link (v, ) 2E, respectively. B. Service model We denote by O the set of information objects or flows that can be captred, processed or transported over the IoT- Clod network, and by P the set of virtal fnctions that can process information objects as part of an offered service. A

4 4 o1 p1 o2 p2 o3 p3 o5 r, d,, z f w f sn, d, o, z q d,o, z tr, d,o, z f v o4 Fig. 2: An example of a service graph, T =(A, O ), with O =5 objects, A =4edges, and P =3virtal fnctions. f pr o, d,o, z f pr i, d,o, z pr given information object o 2Ois, in general, the otpt of a fnction p o 2Pthat reqires the set of objects Z(o) as inpt. For example, in a mobility assistance service, the final object delivered to the end ser may be personalized mobility information reslting from the analysis of data collected by mltiple sensing devices (e.g., GPS devices, wireless sensors, cameras). We represent a generic IoT-Clod service by a directed rooted graph T =(A, O ). For any node o 2O, there is a set of incoming edges {(z,o) 2A : z 2Z(o)}, as shown in Fig. 2. Hence, the set of objects Z(o) Oreqired to generate object o via fnction p o are represented as the children of o in the service graph T. In particlar, the root of the service graph r 2Orepresents the final information object that needs to be delivered to the end ser(s). The service graph hence encodes the relationship between the virtal fnctions that act on the sorce objects to create the final object that needs to be delivered to the end sers. When a ser reqests service, the ser is, in essence, reqesting the final information object or flow represented by the root of the service graph r. In terms of QoS, we se H d,o to denote the maximm delay allowed for the delivery of content object o 2Oat destination d 2V, where we denote by h v and r v the transport delay (in seconds) and reliability (in terms of packet delivery ratio) associated with link (v, ) 2E, respectively, and by h the processing delay at clod node 2V. In addition, we se E (in Watts) to denote the maximm power consmption allowed at battery-powered node 2V in order to garantee its minimm lifetime reqirements. C. Flow model We model the sensing, processing, and transmission of information across the IoT-Clod network in terms of the following ser-object and global flows: User-object flows: are characterized by a triplet (d, o, z), which indicates that the given flow is carrying information of object z 2Osed to deliver final prodct o 2O at destination d 2 V. In particlar, fv tr,d,o,z, fv sn,d,o,z, f pri,d,o,z and f pro,d,o,z indicate the fraction of object z carried/captred/processed by edge (v, ) 2Eor node 2Vfor final prodct o 2Oat destination d 2V, respectively. Note that we differentiate between f pri,d,o,z and f pro,d,o,z, which denote the inpt and otpt flows of the processing nit at node 2V associated with triplet (d, o, z). Fig. 3 illstrates the network flows associated Fig. 3: Inpt/otpt ser-object flows at node 2V. with a given triplet (d, o, z) at node 2V, where pr represents the processing nit that hosts virtal fnctions and q d,o,z is a binary demand parameter that indicates if node 2Vreqests object z 2O. Note that q d,o,z =0if 6= d or z 6= o, since sers only reqest final information objects for themselves. Global flows: fv, tr f sn and f pr determine the total amont of information flow carried, captred, and processed at a given physical link or node, respectively. V. THE INTERNET OF THINGS -CLOUD SERVICE DISTRIBUTION PROBLEM (IOT-CSDP) The IoT-CSDP bilds on the linear CSDP [16] and extends it via the characterization of the access network and the device layer. In particlar, the IoT-CSDP complements the CSDP as follows: The sensing or data acqisition capabilities of end devices, sch as sensors, video cameras or RFID tags, are considered. The capacities and energy costs of processing and transmission resorces are modeled across the entire heterogeneos IoT-Clod network, inclding core, metro, access, and end devices. The limited energy resorces of battery powered devices are taken into accont in order to garantee their minimm lifetime reqirements. The reliability of links is considered in order to characterize the possible packet losses and associated retransmissions, particlarly relevant in low power wireless links. The end-to-end latency is modeled by considering the delay contribtions along the entire service path, from the nodes that generate the sorce data to the delivery of the final agmented information to the end sers. The IoT-CSDP captres the niqe natre of IoT services, typically characterized by a mlticast pstream phase in which sensed data that can be sed for mltiple services and end sers is ploaded to edge clod nodes, and a typically nicast downstream phase in which specific information reslting from the processing of sensed data is delivered in a personalized manner to the end sers. Given that the clod network infrastrctre is shared among mltiple services, the IoT-CSDP assmes a load-

5 5 proportional cost (e.g., energy) model, in which the cost of a given service is proportional to its se of the physical infrastrctre. This reslts in a significantly redced complexity linear program that enables faster reactions to variations of sers service demands. A. Mathematical Formlation The IoT-CSDP is formlated as a minimm cost mixed-cast flow problem as follows: 1) Objective Fnction: We define a generic linear cost fnction characterized by the energy costs of the IoT-Clod network resorces: minimize f tr,f sn,f pr X (v,)2e e tr vfv tr + X (e sn f sn + e pr f pr ) (1) 2V 2) Generalized Flow Conservation Constraints: Userobject flows mst satisfy demand and flow conservation. By modeling the demand q d,o,z as part of the otgoing flow of node, we can se the following generalized flow conservation constraints: f sn,d,o,z q d,o,z + f pri,d,o,z + X + f pro,d,o,z + X v2n () w2n + () f tr,d,o,z w = f tr,d,o,z v 8, d, o, z. (2) Flow conservation constraints state that the otgoing flow associated with a given triplet (d, o, z) mst be eqal to the incoming flow for that same triplet, for any node 2V. As illstrated in Fig. 3, the otgoing flow is composed of the otgoing transport flows, the processing flow leaving node towards the processing nit, and the demand flow; while the incoming flow is composed of the incoming transport flows, the captring flow, and the processing flow going ot of the processing nit. In addition, each processing nit mst satisfy the following flow conservation constraints: f pro,d,o,z apple f pri,d,o,y 8, d, o, z, y 2Z(z). (3) These make sre that in order to have a processed flow z for demand (d, o), a flow associated with each of the inpt objects reqired to generate z, y 2Z(z), for demand (d, o), mst be present at the inpt of the processing nit. 3) Fnction Availability Constraints: These constraints allow restricting the set of virtal fnctions that can be implemented at a given node: f pro,d,o,z =0 8, d, o, z, p z /2P, (4) where P P is the set of virtal fnctions available at node 2V. 4) Sorce Constraints: These constraints initialize the availability of sorce/sensed/captred information objects at their respective sorces: f sn,d,o,z =0 8, d, o, z /2 O, (5) where O S denotes the set of objects that are sorced/captred at node, with S O denoting the set of available sorce information objects. In addition, we need to make sre that sorce objects S Oare not created in the network: f pro,d,o,z =0 8, d, o, z 2S. (6) 5) Mixed-cast Constraints: Mixed-cast constraints allow modeling the nicast or mlticast natre of flows via the corresponding relationship between ser-object and global flows. Since a single captred object can be sed to satisfy mltiple demands, captred flows are said to be mlticast. Hence, the ser-object that captre flows for the same object, bt for different destinations, are allowed to overlap: f sn,d,o,z X z2o apple f sn,z 8, d, o, z, (7) f sn,z B z = f sn 8, (8) where B z denotes the bitrate of object z, and therefore f sn is defined in bps. On the other hand, transport and processing flows can either be nicast or mlticast. In particlar, for IoT services, the transport of captred objects that may be sefl to create mltiple final objects for mltiple sers are modeled as mlticast flows. The mlticast or nicast natre of transport flows that carry processed objects may, however, be modeled as nicast flows if the given processed object is personalized for a specific destination. Processing flows follow the same reasoning and are hence modeled as nicast if the processed flow is destined for a niqe destination, and as mlticast, otherwise. Accordingly, we se the following constraints for mlticast transport and processing flows: fv tr,d,o,z X z2o f pro,d,o,z X z2o apple f tr,z v 8, d, o, z, (9) fv tr,z B z = fv tr0 8, (10) apple f pr,z 8, d, o, z, (11) f pr,z B z z = f pr 8. (12) Note that ser-object flows are sized by the rate in bps of the object, B z. In addition, otpt processing flows are scaled by a factor z that captres possible relative flow size changes associated with the generation of object z from its inpt Z(z) via fnction p z. As a reslt, fv tr0 and f pr are defined in bps. When transporting or processing personalized objects destined for a specific destination, the corresponding transport and processing flows are modeled as nicast. In this case, serobject flows cannot overlap and mst be added across both objects and destinations: X X X fv tr,d,o,z B z = fv tr0 8(v, ), (13) d2v o2o z2o X X X d2v o2o z2o f pro,d,o,z B z z = f pr 8. (14) Note that at this point, transport global flows are denoted by fv tr0, since the actal total transport flows need to accont for possible retransmissions in the presence of nreliable links, as shown in eqation (19) in the following sbsection.

6 6 6) QoS Constraints: Latency: We se d,o,z to denote the (local) delay associated with a particlar ser-object flow (d, o, z). This is compted as the weighted sm of the delay associated with the transport and processing of (d, o, z) flows. That is, 8d, o, z: d,o,z = X (v,)2e f tr,d,o,z v h v + X 2V f pro,d,o,z h, (15) where h v and h (seconds) are the average transport delay at link (v, ) and processing delay at node, respectively. The h v vales inclde the qeing time that each object spends at node v before being forwarded, as well as the propagation time of link (v, ). The h vales depend on the processing capacity of node. Note that in (15), withot loss of generality, we assme inpt processing flows to have zero delay and captre all processing delay with the otpt processing flows, ag as described in [16]. We then se d,o,z to denote the aggregate delay associated with ser-object flow (d, o, z), compted as the sm of the local delay, d,o,z, pls the maximm across the aggregate delays of all of its inpt flows, 8y 2Z(z), as: ag d,o,y ag d,o,z = d,o,z 8d, o, z 2S, (16) d,o,z + ag d,o,y apple ag d,o,z 8d, o, z, y 2Z(z). (17) Finally, the aggregate delay associated with the delivery of final object o at destination d is constrained according to the specified maximm latency reqirements H d,o (in seconds): ag d,o,o apple H d,o 8d, o. (18) Reliability: The reliability of links has a relevant impact on the traffic flow de to the retransmissions cased by packet losses, particlarly in low power wireless links. The average nmber of retransmissions reqired in link (v, ) 2 E is modeled as the reciprocal of its packet delivery ratio: (,w)2e f tr v = f tr0 v /r v 8(v, ). (19) The vale of the packet delivery ratio, r v, is between zero (i.e., disconnected nodes) and 1 (i.e., ideal link), and therefore it increases the actal transport flow. Battery Lifetime: In order to garantee battery-powered devices lifetime reqirements, the total power consmption at node 2 V, considering transport, processing and sensing, mst be lower than its maximm power consmption E (in Watts). Then: X fwe tr tx + X fve tr rx + (v,)2e f sn e sn + f pr e pr apple E 8, (20) where e tx and e rx denote the transmission and reception cost (in Watts per bps), respectively. 7) Capacity Constraints: Global flows mst satisfy capacity constraints: fv tr apple c tr v 8(v, ), (21) apple c sn f sn f pr 8, (22) apple c pr 8. (23) 8) Integer/Fractional Flow Constraints: User-object flows can either be binary or fractional. That is, 8(v, ),,d,o,z: f tr,d,o,z v or f tr,d,o,z v,f pri,d,o,z,f pri,d,o,z,f pro,d,o,z,f sn,d,o,z 2{0, 1}, (24),f pro,d,o,z,f sn,d,o,z 2 [0, 1]. (25) The se of fractional flow variables allows splitting service flows over mltiple paths, providing added flexibility to optimize the se of IoT-Clod resorces. In addition, the reslting linear program admits optimal polynomial time soltions. However, the se of fractional flows challenges the precise comptation of end-to-end latencies in mesh network topologies [35]. In the next section, we choose to solve the IoT-CSDP with binary flow variables in order to precisely model end-to-end delay in mesh network topologies. While the complexity of the reslting integer linear program increases exponentially with the nmber of binary variables, the IoT- CSDP can be efficiently solved to within a few mintes in all tested scenarios. We remark that the soltion to the IoT-CSDP is centralized, i.e., it reqires collecting global information abot service demands and network resorces at a centralized network controller, and disseminating the soltion to all network nodes. In this work, as in [35], we assme that service demands and network resorces change at a longer time-scale than the time needed to implement a new IoT-Clod configration, leaving aspects related to distribted implementations of or soltion to handle faster dynamics as part of ftre work. VI. EVALUATION In this section, we present reslts from the soltion to the IoT-CSDP for illstrative IoT services in smart environments, obtained via the linear programming solver Xpress-MP. We analyze and compare the efficiency of the IoT-Clod soltion, which finds the optimal location of IoT service fnctions exploiting the fll flexibility of the IoT-Clod infrastrctre with: i) a conventional clod approach, in which all service fnctions are centralized at the highest clod layer, ii) the recently proposed clodlet or fog approach [36], in which the processing of IoT services is handled by micro clods located one hop away from the end devices, and iii) a flly distribted approach, in which all service fnctions are exected at the end devices. A. Simlation Details We consider a hierarchical IoT-Clod network architectre composed of three main layers: i) a clod layer, in which clod nodes are organized into 3 tiers, i.e., a head office (HO) node representing the largest centralized data center, intermediate offices (IOs), and end offices (EOs), ii) an

7 7 access layer, composed of base station (BS) nodes hosting micro-clods (MCs) or clodlets [9], and iii) a device layer, containing wireless sensors, smart devices (e.g., smartphones, tablets, smart glasses), and connected vehicles. The specific configration of the IoT-Clod network will be described for each of the simlated scenarios. Precisely modeling the processing capacity (in MIPS) and energy efficiency (in MIPS/W) of the IoT-Clod nodes is difficlt de to the limited information disclosed by clod and network operators. In this paper, we compte approximate vales sing information extracted from [37] (for clod eqipment) and [38] (for wireless sensors/actators), which are presented in Table I. Note that de to resorce consolidation and mltiplexing gains, the efficiency of clod nodes increases at higher layers. Regarding the efficiency of smart devices, given their heterogeneity, we consider a representative range of vales that go from 50 to 1000 MIPS/W [39]. We assme that connected vehicles are eqipped with similar processors [40]. Finally, note that the relatively high processing efficiency of sensors/actators is de to the se of low-power microprocessors. However, they are very limited in terms of battery and compting capacity. In terms of commnication technologies, clod nodes and base stations are connected via optical links, smart devices se 4G or WiFi, and wireless sensors commnicate sing the ZigBee protocol. In order to model ZigBee links, which are particlarly lossy de to their low transmission power, we assme that these are time-varying and follow the logdistance path loss model described in [41] for rban micro-cell scenarios. Then, the path losses (PL) are modeled as follows: PL(dB) =PL 0 (db) log(d) + 3 log(f c /5) + X, (26) where PL 0 is the path loss at the reference distance d 0 (1m), is the path loss exponent, d is the commnication distance, f c is the system freqency (in GHz) and X N 0, 2 is zero mean Gassian noise that models the shadowing effects. This model assmes =2.8 and =4 db [41]. Moreover, a collisionfree MAC layer is assmed to obtain reslts independent from the specific MAC mechanism [42]. Taking into accont the general characteristics of ZigBee transceivers, we assme f c =2.4 GHz, PL 0 (db)=35 db [43], a transmit power of 3 dbm, a sensitivity vale of -91 dbm [44] and omnidirectional antennas. We also consider that the packets generated at wireless sensors have a length of 127 bytes (i.e., standard packet size in IEEE ). On the other hand, we assme that the channel losses in optical, 4G, and WiFi links do not have a relevant impact on their reliability, compared to ZigBee links [41]. The average transport delay vales considered in the simlations inclde the effect of both qeing and propagation delays. While the first is generally higher for the wireless links at the device and access layers de to their limited transmission capacities, the second is typically higher for the optical links connecting clod nodes de to the longer distances among them. Taking this into accont, we assme an average delay of 5 ms for wireless links, while the delay of optical links TABLE I: Typical capacities and efficiencies of IoT-Clod network resorces Capacity Efficiency Clod Node (HO) 53.5 Million MIPS 500 MIPS/W Clod Node (IO) 26 Million MIPS 200 MIPS/W Clod Node (EO) 13 Million MIPS 133 MIPS/W Micro Clod Node (MC) 6.5 Million MIPS 100 MIPS/W Wireless /Actator 1 MIPS 480 MIPS/W Smart Device 2000 MIPS MIPS/W Connected Vehicle 2000 MIPS 1000 MIPS/W Optical Link 4480 Gbps 12.6 nj/bit 4G Link (Down/Up) 72/12 Mbps 76.2/19 µj/bit WiFi Link 150 Mbps 300 nj/bit ZigBee Link 250 kbps 100 nj/bit depends on the hierarchy level of the clod node. In particlar, we assme that a packet needs an additional 2 ms to reach an end office, 7 ms to reach an intermediate office, and 15 ms to reach a head office. The processing delay mainly depends on the particlar processing capabilities of each node. In the simlation reslts, we assme that wireless nodes have an average processing delay of 4 ms, while clod nodes, which can process larger amonts of bits simltaneosly, have a processing delay of 1 ms. In the following, we compte the soltion to the IoT-CSDP in a nmber of representative smart environments (i.e., smart city, smart bilding, and smart mobility) to evalate the impact of different system parameters in the IoT-Clod soltion, sch as transmission and processing efficiencies, nmber of sers, and latency reqirements. B. Smart City We consider the smart city architectre illstrated in Fig. 4a, in which end sers reqest information sing their smart devices. Since smart cities may provide different types of IoT services, we solve the IoT-CSDP for two relevant services: agmented reality and city monitoring. 1) Agmented Reality Service: We consider an agmented reality service in which sers consme video streams that reslt from the combination of the real time video stream captred by the camera of their own smart device, contextal information downloaded from the Internet, and data from the city collected by the WSNs. In particlar, we assme that the agmented video reslts from the combination of 200 kbps of sorce video from the ser device, 50 kbps of city information, and 1 packet/s from each wireless sensor. In terms of processing complexity, we assme that 2000 instrctions per bit are reqired to generate the agmented video. In order to take into accont crrent network access technologies, we assme that half of the sers access the network sing 4G and the other half se a WiFi connexion. Note that both captred and agmented video mst be sent via nicast, bt the contextal information coming from the internet and the wireless sensors can be sent via mlticast, as the same data can be sed for mltiple destinations. The service graph for this application is shown in Fig. 4b. In Fig. 5, we show the average power consmption for different smart device processing efficiencies. The simlation reslts show that the IoT-Clod soltion adapts the offloading

8 8 25 Device Layer Access Layer Clod Layer Smart Devices HO IO IO EO EO EO WSN 1 WSN 2 WSN 3 Virtalized Sensing Platform Average Power Consmption per User (W) Flly Distribted Clodlet Centralized IoT-Clod Processing Efficiency (MIPS/W) Fig. 5: Average power consmption per ser of the agmented reality application for different smart device processing efficiencies. (a) Smart city architectre. The device layer is composed of smart devices, which provide live video streams and reqest agmented videos, and 3 WSNs collecting environmental information arond the city. Measrements Real-Time Video City Information Agmented Video (b) Service graph of the agmented reality application. This combines measrements from wireless sensors, live video streams from the smart devices and city information coming from the HO. Solid and dashed arrows indicate mlticast and nicast flows, respectively. Measrements (WSN1) Measrements (WSN2) Measrements (WSN3) Aggregated Measrements City Information A City Information B City Information C (c) Service graph of the city monitoring application. The city information is obtained throgh the analysis of the aggregated measrements collected by 3 different WSNs. readings are allowed to be sent via mlticast, bt the city information can only be sent via mlticast for mltiple sers reqesting the same information simltaneosly. Fig. 4: Smart city environment decisions according to the processing efficiency of end devices. Specifically, note that if the device processing efficiency is sfficiently high, i.e., 500 MIPS/W, the IoT-Clod soltion allocates the agmented video processing fnction at the end devices. Otherwise, some processing tasks have to be offloaded to the clod in order to redce the overall processing cost. We show that for efficiencies between 50 and 300 MIPS/W the IoT-Clod ses hybrid soltions, in which the placement strategy is not niform for all sers. In these cases, the processing tasks of WiFi sers are offloaded to the Clod, while the network access cost of 4G devices keeps their video processing fnctionality at the end devices. We can also observe that the clodlet soltion yields an overall power consmption that is even higher than that of the centralized approach. This is becase the micro clod nodes have lower processing efficiencies than the larger clod data centers, and this does not get compensated by the savings in terms of transport cost. Note that we consider the se of highly efficient optical fiber links throghot the clod layer. 2) City Monitoring Service: Thanks to the deployment of large scale wireless sensor networks, smart cities can provide real-time information abot the city and its infrastrctres. In this section, we consider a city monitoring service in which end sers reqest city information that is generated from the analysis of data collected by wireless sensors, sch as in [45]. The service graph of this application is shown in Fig. 4c. We evalate the impact of the city information bitrate and the nmber of simltaneos sers reqesting the same information. In Fig. 6, we compare the power consmption for different ser demands in terms of bitrate. A high bitrate means that the ser reqests detailed information, while low bitrates indicate that the service only provides the most relevant information. We assme that end sers reqest City Information B (See Fig. 4c), which receives the aggregated measrements of WSN 1, WSN 2 and WSN 3 as inpt. We assme smart devices have a processing efficiency of 1000 MIPS/W. The processing complexity of analyzing the data from each WSN is 500 instrctions per bit, while the complexity of generating

9 9 Average Power Consmption per User (W) Flly Distribted Clodlet Centralized IoT-Clod Personalized Information Rate (kbps) Fig. 6: Average power consmption per ser of a city monitoring application for different personalized information rates. Average Power Consmption (W) Flly Distribted Clodlet Centralized IoT-Clod Simltaneos Wifi Users Flly Distribted Clodlet Centralized IoT-Clod (a) Simltaneos WiFi sers the personalized information reqested by the ser is 5000 instrctions per bit. An average rate of 1 packet per second is generated by each wireless sensor. We consider the same nmber of WiFi and 4G sers, and an average nmber of 4 sers simltaneosly reqesting the same information in total. Note that the sensor readings can be sent via mlticast, since they can be sed to generate different final information objects. However, personalized information can be sent via mlticast only when mltiple sers reqest the same information at the same time. Otherwise, these mst be sent via nicast. Nevertheless, we assme that the transmission from the base stations to the end sers is always nicast, as neither WiFi nor 4G mlticast technologies are yet commercially available. As we can observe, if sers reqest a low bitrate the IoT-Clod soltion tends to consolidate the processing fnctionalities at the HO in order to redce the total processing cost. As a reslt, the overall power consmption is redced by more than 84% compared to the flly distribted soltion. However, the centralized clod approach becomes less efficient at high bitrates de to the transport cost from the clod nodes to the smart devices, particlarly to the 4G devices. Then, the IoT- Clod soltion pshes some processing fnctionalities closer to the sers in order to redce transport costs, at the expense of redcing the consolidation gain. In particlar, the IoT-Clod soltion moves the fnction that generates the personalized data to the end devices for bitrates above 20 kbps, while keeping the sensor data analysis at the clod layer. We can observe that the overall power consmption of the centralized and clodlet soltions are severely affected by the reqested bitrate, being their cost more than 78% higher than the cost of the IoT-Clod soltion if the bitrate exceeds 200 kbps. In Fig. 7, we compare the power consmption for different average nmber of sers reqesting simltaneosly the same information. In this case, we separate the power consmption of WiFi and 4G sers to observe the impact of simltaneos sers in each case. The bitrate of the city information is assmed to be 200 kbps and the processing efficiency of the smart devices is 1000 MIPS/W. In Fig. 7a, we observe that if there are more than 2 simltaneos WiFi sers, it is Average Power Consmption (W) Simltaneos 4G Users (b) Simltaneos 4G sers Fig. 7: Average power consmption per ser of a city monitoring service for different nmber of simltaneos sers. more energy efficient to consolidate the processing tasks at the HO than to process them locally, even if this increases the total transport cost. Note that if the information reqested is generated at the HO, the network can take more advantage of mlticast delivery. However, for sers accessing the network sing 4G (see Fig. 7b), the additional transport cost of 4G networks cannot be compensated even with 4 simltaneos 4G sers, being the overall cost of the centralized soltion 80% higher than that of the IoT-Clod soltion. Nevertheless, it is important to note that if wireless mlticast streaming schemes were considered, the impact of processing consolidation wold be more relevant in both cases. We again observe the higher power consmption of the clodlet soltion de to its low processing efficiency. This is particlarly relevant in Fig. 7a de to the higher significance of processing costs compared to transport costs in WiFi networks. We remark that the effect of the nmber of sers is specially relevant nder the presence of mlticast flows. Note that if all flows are nicast, then, nder the adopted linear cost model, the total power consmption wold be proportional to the total nmber of sers, since each transport/processing fnction needs to be condcted individally for each end ser.

10 10 C. Smart Bilding In this section, we consider an IoT service that controls a bilding atomation system. A WSN that collects 5 different types of measrements (e.g., hmidity, temperatre, light, smoke, occpancy) is assisted by the Clod to manage 5 actators distribted across the bilding (See Fig. 8a). According to the sensor readings, and after a simple data analysis, the actators manage different atonomos systems. In terms of processing complexity, both data aggregation and data analysis reqire 50 instrctions per bit. We also assme that the measrements from wireless sensors can be combined sing data aggregation techniqes as long as they contain the same type of information. This redces traffic by a compression rate that depends on the spatial and temporal correlation of data. A realistic compression rate of 50% [46] is assmed in these simlations. The sensor readings may be sent via mlticast, bt the actator decisions are sent via nicast. However, note that in this case the distribtion mode does not have any impact in the total traffic flow, since every object has only one possible destination, and hence there are no mlticast opportnities. The service graph associated with this service is shown in Fig.8b. In order to observe the impact of the limited battery capacity of wireless actators, we constrain their average power consmption according to their expected battery lifetime. Note that battery replacement is a costly operation and therefore nodes shold not deplete their batteries before their schedled maintenance. With this in mind, we only allow soltions that assre a minimm lifetime of one year for the wireless actators. Assming that each actator is eqipped with a battery storing 27 kj [47], this translates into a maximm power of E = 856µW. In Fig. 9, we compare the power consmption for different packet generation rates at the wireless sensors. A high bitrate increases the accracy of the service, bt it also increases the power consmption of the wireless sensors and actators. Note that when the average power consmption of a given soltion is not shown, it indicates that sch soltion is infeasible, i.e., it violates QoS or capacity constraints. In scenarios that reqire more than 2 packets/s from each sensor, the processing tasks are flly offloaded to the Clod to redce the power consmption of the wireless actators. However, if the packet rate can be redced, the limited processing capacity of actators can be sed to redce the traffic sent to the Clod. As we can observe, the device layer can control the actators withot the assistance of the clod infrastrctre as long as the packet rate is lower than 1.5 packets/s. As a reslt, the total transport consmption is redced. However, the processing tasks have to be offloaded to the Clod at higher rates in order to satisfy the minimm lifetime reqirements of the wireless actators. Therefore, in this case, the IoT-Clod soltion corresponds to the centralized soltion. It is worth noting that the optimized IoT-Clod soltion only ses the clod resorces when they are strictly necessary. Device Layer Access Layer Clod Layer IO HO EO EO EO IO Virtalized Network + Actator (a) Smart bilding architectre. The device layer is composed of wireless sensors and actators. The sensors collect the environmental measrements, while the actators control the bilding atomation system. Average Power Consmption (W) Measrements A Measrements B Measrements E Aggregated Measrements Actator A Decision Actator B Decision Actator E Decision (b) Service graph of the smart bilding application. The sensor readings are first aggregated, to redce the transport flow, and then analyzed to control the bilding atomation system sing the wireless actators. Solid and dashed arrows state for mlticast and nicast flows, respectively Fig. 8: Smart bilding environment Flly Distribted Clodlet Centralized IoT-Clod Packets/s Fig. 9: Average power consmption of an atonomos sensing actation service for different packet generation rates. Note that missing vales indicate that the soltion is infeasible.

11 Device Layer Access Layer Clod Layer IO HO EO EO EO (a) Smart traffic architectre. The device layer is composed by the connected vehicles, which provide life video streams to the traffic management system. IO Average Power Consmption (W) Flly Distribted Clodlet Centralized IoT-Clod Maximm Latency (ms) Fig. 11: Average power consmption of a traffic management service for different latency reqirements. Note that missing vales indicate that the soltion is infeasible. Video Streams from Vehicles Traffic Lights Measrements Combined Video Analyzed Measrements Traffic Decisions (b) Service graph of the smart mobility application. The measrements collected at the traffic lights are combined with the video streams of vehicles. Solid and dashed arrows indicate mlticast and nicast flows, respectively. D. Smart Mobility Fig. 10: Smart mobility environment In this section, we consider a service that interconnects city vehicles among each other, as well as with city traffic lights. This enables optimizing city traffic, redcing congestion, and increasing road safety. We assme the scenario in Fig. 10a, in which vehicles are connected to other nearby vehicles and to their closest base stations. In addition, each base station has 5 traffic lights connected to it, each generating 127 bytes per second of traffic information. The vehicles generate video streams of 200 kbps. The streams from nearby vehicles are combined and compressed sing a scaling factor of 0.5. The measrements from the traffic lights and the video streams are then jointly analyzed to predict traffic jams and manage traffic lights accordingly. In particlar, we consider the service graph shown in Fig. 10b. In terms of processing complexity, we assme that processing the traffic measrements reqires 50 instrctions per bit, the combination of the video streams 2000 instrctions per bit, and the generation of the traffic decisions reqires 5000 instrctions per bit. We frther assme that vehicles embedded compting resorces have a processing efficiency of 1000 MIPS/W. In Fig. 11, we show the average power consmption for different latency reqirements. Low end-to-end service latencies are specially relevant for smart mobility services in which processed information is sed to coordinate city traffic in real- time. It is important to remark that there exists a minimm end-to-end service latency dictated by the network topology and the adopted soltion approach. As shown in Fig. 10a, sorce streams need to be sent at least to the base stations (5 ms), which control the traffic lights, then forwarded to be jointly processed (i.e., at least 4 ms), and finally otpt decisions mst be sent back to the base stations (i.e., at least 2 ms) that control the traffic lights. Hence, the minimm achievable transport delay is 11 ms. In addition, we need to consider the processing delay, which is at least 2 ms if the processing tasks are placed at a clod location (i.e., 1 ms for each processing fnction). Note that this minimm endto-end latency can only be achieved with the IoT-Clod and clodlet soltions. A flly distribted soltion can only achieve a minimm latency of 28 ms (i.e., 20 ms for transport and 8 ms for processing). Finally, the centralized clod soltion has a minimm latency of 55 ms (i.e., 53 ms for transport and 2 ms for processing). This can be observed in Fig. 11, where maximm latency reqirements below 28 ms can only be satisfied by the optimized IoT-Clod soltion. It is worth noting that the IoT-Clod optimal soltion does not need to place the virtal fnctions at the access layer to achieve its minimm latency. Since the information processed at the micro clod nodes needs to go throgh the clod layer for sharing prposes, placing the virtal fnctions at the edge of the clod layer (EOs) redces the processing cost withot increasing the end-to-end latency. The reslts presented in this paper clearly show the benefits that arise from the flexibility of an optimized IoT- Clod soltion and the limitations of crrent centralized, flly distribted, and clodlet soltions for the efficient delivery of next generation IoT services. On one hand, centralizing all processing tasks at a single core-level data center is shown to be efficient in terms of processing costs, while incrring transport costs and end-to-end latencies that become nsstainable for increasingly bandwidth-intensive lowlatency IoT services. On the other hand, placing all processing

12 12 fnctionality at the device layer, while appropriate for low complexity operations, can seriosly compromise the desirable lightweight, long battery life properties of end devices. An intermediate soltion, sch as placing the processing tasks at the access layer, improves service latencies compared to a centralized soltion, bt it also redces overall processing efficiency de to the lower compting capabilities of micro clod nodes. In contrast, we have shown that an optimized IoT- Clod soltion overcomes these limitations by placing service fnctions at different layers of the IoT-Clod infrastrctre as a fnction of the inpt parameters, striking the right balance between processing efficiency, transport efficiency, and endto-end latency. VII. CONCLUSIONS The conflence of distribted clod networking and the Internet of Things (IoT) enables a new class of services that create agmented information from the clod analysis of IoT data. IoT-Clod networks redce the distance between end sers and clod resorces sing edge clod nodes distribted across the network, in order to spport the key low latency, mobility, and location-awareness reqirements of IoT services. We proposed a comprehensive optimization framework to enhance the efficiency of IoT services in next generation smart environments. We formlated the service distribtion problem (SDP) in IoT-Clod networks (IoT-CSDP) as a mincost mixed-cast flow problem. The soltion to the IoT-CSDP determines the optimal placement of service fnctions and the roting of network flows, taking into accont the heterogeneos capacities and efficiencies of sensing, compting, and transport resorces across the distribted IoT-Clod infrastrctre. We solved the IoT-CSDP for the optimization of IoT services in mltiple smart environments. Reslts show that the IoT-CSDP soltion captres the critical tradeoffs that appear in IoT-Clod platforms de to the heterogeneity of IoT services, clod network technologies, and end ser devices. When compared to crrent soltions, smart IoT services optimized over a flly virtalized IoT-Clod platform are shown to garantee stringent QoS reqirements in terms of reliability, battery lifetime, and end-to-end latency, while redcing overall power consmption by more than 80%. Motivated by these promising reslts, interesting directions for ftre work inclde implementing the proposed soltion in a clod compting testbed, designing approximation algorithms of improved comptational complexity, and stdying the benefit of distribted optimization techniqes to enable local reactions to fast system dynamics. VIII. ACKNOWLEDGEMENTS This work was spported in part by the Spanish Government nder project TEC R. REFERENCES [1] L. Atzori, A. Iera, and G. Morabito, The internet of things: A srvey, Compter Networks, vol. 54, no. 15, pp , [2] A. Al-Fqaha, M. Gizani, M. Mohammadi, M. Aledhari, and M. Ayyash, Internet of things: A srvey on enabling technologies, protocols, and applications, Commnications Srveys Ttorials, IEEE, vol. 17, no. 4, pp , Forthqarter [3] C. Zh, V. C. M. Leng, L. Sh, and E. C. H. Ngai, Green internet of things for smart world, IEEE Access, vol. 3, pp , [4] N. Alhakbani, M. M. Hassan, M. A. Hossain, and M. Alnem, Internet and Distribted Compting Systems: 7th International Conference, IDCS 2014, Calabria, Italy, September 22-24, Proceedings. Cham: Springer International Pblishing, 2014, ch. A Framework of Adaptive Interaction Spport in Clod-Based Internet of Things (IoT) Environment, pp [5] K. S, J. Li, and H. F, Smart city and the applications, in Electronics, Commnications and Control (ICECC), 2011 International Conference on, Sept 2011, pp [6] C. Perera, C. H. Li, and S. Jayawardena, The emerging internet of things marketplace from an indstrial perspective: A srvey, IEEE Transactions on Emerging Topics in Compting, vol. 3, no. 4, pp , Dec [7] S. Sarkar, S. Chatterjee, and S. Misra, Assessment of the sitability of fog compting in the context of internet of things, Clod Compting, IEEE Transactions on, vol. PP, no. 99, pp. 1 1, [8] DennisBoley, Estimating a data center s electrical carbon footprint, White Paper 66, [9] M. Satyanarayanan, P. Bahl, R. Caceres, and N. Davies, The case for VM-based clodlets in mobile compting, Pervasive Compting, IEEE, vol. 8, no. 4, pp , Oct [10] W. Shiqiang, et al, Mobile micro-clod: Application classification, mapping, and deployment, in Annal Fall Meeting of ITA (AMITA), [11] F. Bonomi, R. Milito, J. Zh, and S. Addepalli, Fog compting and its role in the internet of things, in Proceedings of the First Edition of the MCC Workshop on Mobile Clod Compting, ser. MCC 12. New York, NY, USA: ACM, 2012, pp [12] H. Chang, A. Hari, S. Mkherjee, and T. Lakshman, Bringing the clod to the edge, in Compter Commnications Workshops (INFOCOM WKSHPS), 2014 IEEE Conference on, April 2014, pp [13] B. Addis, D. Ardagna, B. Paniccci, M. Sqillante, and L. Zhang, A hierarchical approach for the resorce management of very large clod platforms, Dependable and Secre Compting, IEEE Transactions on, vol. 10, no. 5, pp , Sept [14] L. Lei, Z. Zhong, K. Zheng, J. Chen, and H. Meng, Challenges on wireless heterogeneos networks for mobile clod compting, Wireless Commnications, IEEE, vol. 20, no. 3, pp , Jne [15] S. Nastic, S. Sehic, D.-H. Le, H.-L. Trong, and S. Dstdar, Provisioning software-defined IoT clod systems, in Ftre Internet of Things and Clod (FiClod), 2014 International Conference on, Ag 2014, pp [16] M. Barcelo, J. Llorca, A. Tlino, and N. Raman, The clod service distribtion problem in distribted clod networks, in IEEE ICC 2015 SAC - Data Storage and Clod Compting, London, United Kingdom, Jn [17] I. Stojmenovic and S. Wen, The fog compting paradigm: Scenarios and secrity isses, in Compter Science and Information Systems (FedCSIS), 2014 Federated Conference on, Sept 2014, pp [18] A. Alamri, W. S. Ansari, M. M. Hassan, M. S. Hossain, A. Alelaiwi, and M. A. Hossain, A srvey on sensor-clod: Architectre, applications, and approaches, International Jornal of Distribted Networks, vol. 2013, no , [19] S. Misra, S. Chatterjee, and M. Obaidat, On theoretical modeling of sensor clod: A paradigm shift from wireless sensor network, Systems Jornal, IEEE, vol. PP, no. 99, pp. 1 10, [20] N. Mitton, S. Papavassilio, A. Pliafito, and K. Trivedi, Combining clod and sensors in a smart city environment, EURASIP Jornal on Wireless Commnications and Networking, vol. 2012, no. 1, [21] M. Aslam, S. Rea, and D. Pesch, Service provisioning for the WSN clod, in 2012 IEEE 5th International Conference on Clod Compting, Jne 2012, pp [22] C. Zh, H. Wang, X. Li, L. Sh, L. T. Yang, and V. C. M. Leng, A novel sensory data processing framework to integrate sensor networks with mobile clod, IEEE Systems Jornal, vol. PP, no. 99, pp. 1 12, [23] C. Zh, Z. Sheng, V. C. M. Leng, L. Sh, and L. T. Yang, Toward offering more sefl data reliably to mobile clod from wireless sensor network, IEEE Transactions on Emerging Topics in Compting, vol. 3, no. 1, pp , [24] C. Zh, H. Nicanfar, V. C. M. Leng, and L. T. Yang, An athenticated trst and reptation calclation and management system for clod and sensor networks integration, IEEE Transactions on Information Forensics and Secrity, vol. 10, no. 1, pp , Jan 2015.

13 13 [25] C. Zh, V. C. M. Leng, L. T. Yang, X. H, and L. Sh, Collaborative location-based sleep schedling to integrate wireless sensor networks with mobile clod compting, in 2013 IEEE Globecom Workshops (GC Wkshps), Dec 2013, pp [26] C. H. Li, J. Fan, J. W. Branch, and K. K. Leng, Toward QoI and energy-efficiency in internet-of-things sensory environments, IEEE Transactions on Emerging Topics in Compting, vol. 2, no. 4, pp , Dec [27] A. Mkherjee, H. Pal, S. Dey, and A. Banerjee, ANGELS for distribted analytics in IoT, in Internet of Things (WF-IoT), 2014 IEEE World Form on, March 2014, pp [28] N. Trong, G. M. Lee, and Y. Ghamri-Dodane, Software defined networking-based vehiclar adhoc network with fog compting, in Integrated Network Management (IM), 2015 IFIP/IEEE International Symposim on, May 2015, pp [29] S. Sarkar and S. Misra, Theoretical modelling of fog compting: a green compting paradigm to spport IoT applications, IET Networks, vol. 5, no. 2, pp , [30] I. Farris, L. Militano, M. Nitti, L. Atzori, and A. Iera, Federated edgeassisted mobile clods for service provisioning in heterogeneos IoT environments, in Internet of Things (WF-IoT), 2015 IEEE 2nd World Form on, Dec 2015, pp [31] M. Aazam and E.-N. Hh, Fog compting micro datacenter based dynamic resorce estimation and pricing model for IoT, in Advanced Information Networking and Applications (AINA), 2015 IEEE 29th International Conference on, March 2015, pp [32] M. Aazam, M. St-Hilaire, C. H. Lng, and I. Lambadaris, PRE-Fog: IoT trace based probabilistic resorce estimation at fog, in th IEEE Annal Consmer Commnications Networking Conference (CCNC), Jan 2016, pp [33] J. Bngo, Embedded systems programming in the clod: A novel approach for academia, Potentials, IEEE, vol. 30, no. 1, pp , Jan [34] The programmable clod network - a primer on SDN and NFV, Bell Labs Strategic White Paper, Jne [35] J. Llorca and A. Tlino, The content distribtion problem and its complexity classification, Alcatel-Lcent, Tech. Rep., [36] M. Satyanarayanan, R. Schster, M. Ebling, G. Fettweis, H. Flinck, K. Joshi, and K. Sabnani, An open ecosystem for mobile-clod convergence, Commnications Magazine, IEEE, vol. 53, no. 3, pp , March [37] A. Vishwanath, F. Jalali, K. Hinton, T. Alpcan, R. Ayre, and R. Tcker, Energy consmption comparison of interactive clod-based and local applications, Selected Areas in Commnications, IEEE Jornal on, vol. 33, no. 4, pp , April [38] M. Ilyas and I. Mahgob, Smart Dst: Network Applications, Architectre and Design. CRC Press, [39] P. Zheng and L. Ni, Smart Phone and Next Generation Mobile Compting. Morgan Kafmann, [40] Qalcomm atomotive infotainment. [Online]. Available: [41] P. Kyösti, J. Meinilä, L. Hentilä, X. Zhao, and T. Jämsä, WINNER II channel models, EBITG, TUI, UOULU, CU/CRC, NOKIA, Tech. Rep., [42] V. G. Gimaraes, A. Bachspiess, and R. M. de Moraes, Dynamic timed energy efficient and data collision free MAC protocol for wireless sensor networks, IEEE Latin America Transactions, vol. 13, no. 2, pp , Feb [43] A. Martinez-Sala, J.-M. Molina-Garcia-Pardo, E. Egea-Ldpez, J. Vales- Alonso, L. Jan-Llacer, and J. Garcia-Haro, An accrate radio channel model for wireless sensor networks simlation, Commnications and Networks, Jornal of, vol. 7, no. 4, pp , Dec [44] Atmel, AT86RF230: Low power 2.4 GHz transceiver for ZigBee, IEEE , 6LoWPAN, RF4CE and ISM applications, [45] C. Csiszar and D. Foldes, Analysis and modelling methods of rban integrated information system of transportation, in Smart Cities Symposim Prage (SCSP), 2015, Jne 2015, pp [46] C. Cappiello and F. A. Schreiber, Experiments and Analysis of Qalityand Energy-Aware Data Aggregation Approaches in WSNs, in QDB 2012, 10th International Workshop on Qality in Databases, [47] A. Sivagami, K. Pavai, D. Sridharan, and S. SatyaMrty, Estimating the energy consmption of wireless sensor node: Iris, International Jornal on Recent Trends in Engineering and Technology, vol. 3, no. 4, pp , May Marc Barcelo received his M.Sc. degree in electrical engineering (Hons.) and Ph.D. degree (cm lade) from Universitat Atonoma de Barcelona (UAB) in 2010 and 2015, respectively. From 2009 to 2015 he was with the UAB s Telecommnications and Systems Engineering Department participating in several international R&D projects, sch as ADIBEAM and SINTONIA. Dring 2014, he was an International Research Intern at Nokia Bell Labs, New Jersey, USA. He joined Ikerlan, Mondragon-Arrasate, Spain, in 2016, where he is crrently a Research Scientist in the Cybersecre IoT department. He has pblished over 20 papers in recognized international jornals and conferences. Crrently, his research efforts are focsed on the design of efficient commnication architectres for the Internet of Things. In 2011, he received the COIT/AEIT prize for his M.Sc. thesis. Alejandro Correa received his B.Sc and M.Sc. in Electrical Engineering (with honors) from Universitat Atonoma de Barcelona (UAB), in 2008 and 2010, respectively. Since 2011 he is a PhD candidate at the Telecommnications and Systems Engineering Department of UAB where he has been involved in several R&D projects, sch as ATLANTIDA and 3SENS. He has pblished in 8 international jornals and 7 international conferences. His main research interests inclde pedestrian indoor positioning systems and wireless sensor networks. Jaime Llorca received the B.E. degree in Electrical Engineering from the Universitat Politecnica de Catalnya (UPC), Barcelona, Spain, in 2001, and the M.S. and Ph.D. degrees in Electrical and Compter Engineering from the University of Maryland, College Park, MD, USA, in 2003 and 2008, respectively. He held a post-doctoral position at the Center for Networking of Infrastrctre s (CNIS), College Park, MD, USA, from 2008 to He joined Nokia Bell Labs at Holmdel, NJ, USA, in 2010, where he is crrently a Research Scientist in the Network Algorithms Department. His research interests inclde energy efficient networks, distribted clod networking, content distribtion, resorce allocation, network information theory, and network optimization. He is a recipient of the 2007 Best Paper Award at the IEEE International Conference on s, Networks and Information Processing (ISSNIP), the 2016 Best Paper Award at the IEEE International Conference on Commnications (ICC), and the 2015 Jimmy H.C. Lin Award for Innovation.

14 14 Antonia M. Tlino (S 00 M 03 SM 05 F 13) received the Ph.D. degree in Electrical Engineering from Seconda Universita degli Stdi di Napoli, Italy, in She held research positions at Princeton University, at the Center for Wireless Commnications, Ol, Finland and at Universita degli Stdi del Sannio, Benevento, Italy. From 2002 she has joined the Faclty of the Universita degli Stdi di Napoli Federico II, and in 2009 she joined Bell Labs. From 2011, Dr. Tlino is Member of the Editorial Board of the IEEE Transactions on Information Theory and in 2013, she was elevated to IEEE Fellow. She has received several paper awards and among the others the 2009 Stephen O. Rice Prize in the Field of Commnications Theory for the best paper pblished in the IEEE TRANSACTION ON COMMUNICATION in She has been principal investigator of several research projects sponsored by the Eropean Union and the Italian National Concil, and was selected by the National Academy of Engineering for the Frontiers of Engineering program in Her research interests lie in the area of commnication systems approached with the complementary tools provided by signal processing, information theory and random matrix theory. Prof. J.L. Vicario, received both the degree in electrical engineering and the Ph.D. degree (cm lade) from the Universitat Politecnica de Catalnya (UPC), Barcelona, in 2002 and 2006, respectively. He also holds an MBA from the IESE Bsiness School, Universidad de Navarra. Since September 2006, he is an Associate Professor at the Universitat Atonoma de Barcelona teaching corses in digital commnications and signal processing and participating in several R&D Projects sch as IST FP6 SATNEX, ESA DINGPOS, ESA ADIBEAM, CENIT ATLANTIDA, CENIT SINTONIA (as PI), PROFIT INTERRURAL (as PI), INFOREGIO XALOC (as PI), AVANZA DESSO (as PI) and RECERCAIXA 3SENS. He has a wide expertise on statistical signal processing, pblishing 24 papers in recognized international jornals and arond 50 papers in conferences. Crrently, his research efforts are focsed on the development of data mining algorithms covering different areas sch as wireless sensor networks, twitter analysis, financial applications and air traffic management. Antoni Morell received the M.Sc. degree in electrical and electronic engineering and the Ph.D. degree from Universitat Politecnica de Catalnya (UPC), Barcelona, in 2002 and 2008, respectively. He was with the Signal Theory and Commnications Department, UPC, from 2002 to He joined Universitat Atonoma de Barcelona in 2005, where he crrently holds an Associate Professor position. He teaches corses on commnications, signals and systems. He has been involved in over 10 R&D projects, national and international, being the Principal Investigator in two of them. He has expertise in optimization techniqes applied to commnications, also in the field of wireless sensor networks. He has pblished over 40 papers in recognized international jornals and conference proceedings and his crrent research interests inclde smart data applications of machine learning as well as inertial-aided indoor positioning and roting/access strategies for WSNs.

On the Computational Complexity and Effectiveness of N-hub Shortest-Path Routing

On the Computational Complexity and Effectiveness of N-hub Shortest-Path Routing 1 On the Comptational Complexity and Effectiveness of N-hb Shortest-Path Roting Reven Cohen Gabi Nakibli Dept. of Compter Sciences Technion Israel Abstract In this paper we stdy the comptational complexity

More information

Networks An introduction to microcomputer networking concepts

Networks An introduction to microcomputer networking concepts Behavior Research Methods& Instrmentation 1978, Vol 10 (4),522-526 Networks An introdction to microcompter networking concepts RALPH WALLACE and RICHARD N. JOHNSON GA TX, Chicago, Illinois60648 and JAMES

More information

An Adaptive Strategy for Maximizing Throughput in MAC layer Wireless Multicast

An Adaptive Strategy for Maximizing Throughput in MAC layer Wireless Multicast University of Pennsylvania ScholarlyCommons Departmental Papers (ESE) Department of Electrical & Systems Engineering May 24 An Adaptive Strategy for Maximizing Throghpt in MAC layer Wireless Mlticast Prasanna

More information

What s New in AppSense Management Suite Version 7.0?

What s New in AppSense Management Suite Version 7.0? What s New in AMS V7.0 What s New in AppSense Management Site Version 7.0? AppSense Management Site Version 7.0 is the latest version of the AppSense prodct range and comprises three prodct components,

More information

Cost Based Local Forwarding Transmission Schemes for Two-hop Cellular Networks

Cost Based Local Forwarding Transmission Schemes for Two-hop Cellular Networks Cost Based Local Forwarding Transmission Schemes for Two-hop Celllar Networks Zhenggang Zhao, Xming Fang, Yan Long, Xiaopeng H, Ye Zhao Key Lab of Information Coding & Transmission Sothwest Jiaotong University,

More information

REPLICATION IN BANDWIDTH-SYMMETRIC BITTORRENT NETWORKS. M. Meulpolder, D.H.J. Epema, H.J. Sips

REPLICATION IN BANDWIDTH-SYMMETRIC BITTORRENT NETWORKS. M. Meulpolder, D.H.J. Epema, H.J. Sips REPLICATION IN BANDWIDTH-SYMMETRIC BITTORRENT NETWORKS M. Melpolder, D.H.J. Epema, H.J. Sips Parallel and Distribted Systems Grop Department of Compter Science, Delft University of Technology, the Netherlands

More information

A Hybrid Weight-Based Clustering Algorithm for Wireless Sensor Networks

A Hybrid Weight-Based Clustering Algorithm for Wireless Sensor Networks Open Access Library Jornal A Hybrid Weight-Based Clstering Algorithm for Wireless Sensor Networks Cheikh Sidy Mohamed Cisse, Cheikh Sarr * Faclty of Science and Technology, University of Thies, Thies,

More information

The Disciplined Flood Protocol in Sensor Networks

The Disciplined Flood Protocol in Sensor Networks The Disciplined Flood Protocol in Sensor Networks Yong-ri Choi and Mohamed G. Goda Department of Compter Sciences The University of Texas at Astin, U.S.A. fyrchoi, godag@cs.texas.ed Hssein M. Abdel-Wahab

More information

Energy-Efficient Mobile-Edge Computation Offloading for Applications with Shared Data

Energy-Efficient Mobile-Edge Computation Offloading for Applications with Shared Data Energy-Efficient Mobile-Edge Comptation Offloading for Applications with Shared Data Xiangy He, Hong Xing, Ye Chen, Armgam Nallanathan School of Electronic Engineering and Compter Science Qeen Mary University

More information

Minimal Edge Addition for Network Controllability

Minimal Edge Addition for Network Controllability This article has been accepted for pblication in a ftre isse of this jornal, bt has not been flly edited. Content may change prior to final pblication. Citation information: DOI 10.1109/TCNS.2018.2814841,

More information

Image Compression Compression Fundamentals

Image Compression Compression Fundamentals Compression Fndamentals Data compression refers to the process of redcing the amont of data reqired to represent given qantity of information. Note that data and information are not the same. Data refers

More information

Cautionary Aspects of Cross Layer Design: Context, Architecture and Interactions

Cautionary Aspects of Cross Layer Design: Context, Architecture and Interactions Cationary Aspects of Cross Layer Design: Context, Architectre and Interactions Vikas Kawadia and P. R. Kmar Dept. of Electrical and Compter Engineering, and Coordinated Science Lab University of Illinois,

More information

Isilon InsightIQ. Version 2.5. User Guide

Isilon InsightIQ. Version 2.5. User Guide Isilon InsightIQ Version 2.5 User Gide Pblished March, 2014 Copyright 2010-2014 EMC Corporation. All rights reserved. EMC believes the information in this pblication is accrate as of its pblication date.

More information

Requirements Engineering. Objectives. System requirements. Types of requirements. FAQS about requirements. Requirements problems

Requirements Engineering. Objectives. System requirements. Types of requirements. FAQS about requirements. Requirements problems Reqirements Engineering Objectives An introdction to reqirements Gerald Kotonya and Ian Sommerville To introdce the notion of system reqirements and the reqirements process. To explain how reqirements

More information

NETWORK PRESERVATION THROUGH A TOPOLOGY CONTROL ALGORITHM FOR WIRELESS MESH NETWORKS

NETWORK PRESERVATION THROUGH A TOPOLOGY CONTROL ALGORITHM FOR WIRELESS MESH NETWORKS ETWORK PRESERVATIO THROUGH A TOPOLOGY COTROL ALGORITHM FOR WIRELESS MESH ETWORKS F. O. Aron, T. O. Olwal, A. Krien, Y. Hamam Tshwane University of Technology, Pretoria, Soth Africa. Dept of the French

More information

Unit Testing with VectorCAST and AUTOSAR

Unit Testing with VectorCAST and AUTOSAR Unit Testing with VectorCAST and AUTOSAR Vector TechDay Software Testing with VectorCAST V1.0 2018-11-15 Agenda Introdction Unit Testing Demo Working with AUTOSAR Generated Code Unit Testing AUTOSAR SWCs

More information

A sufficient condition for spiral cone beam long object imaging via backprojection

A sufficient condition for spiral cone beam long object imaging via backprojection A sfficient condition for spiral cone beam long object imaging via backprojection K. C. Tam Siemens Corporate Research, Inc., Princeton, NJ, USA Abstract The response of a point object in cone beam spiral

More information

Statistical Methods in functional MRI. Standard Analysis. Data Processing Pipeline. Multiple Comparisons Problem. Multiple Comparisons Problem

Statistical Methods in functional MRI. Standard Analysis. Data Processing Pipeline. Multiple Comparisons Problem. Multiple Comparisons Problem Statistical Methods in fnctional MRI Lectre 7: Mltiple Comparisons 04/3/13 Martin Lindqist Department of Biostatistics Johns Hopkins University Data Processing Pipeline Standard Analysis Data Acqisition

More information

The final datapath. M u x. Add. 4 Add. Shift left 2. PCSrc. RegWrite. MemToR. MemWrite. Read data 1 I [25-21] Instruction. Read. register 1 Read.

The final datapath. M u x. Add. 4 Add. Shift left 2. PCSrc. RegWrite. MemToR. MemWrite. Read data 1 I [25-21] Instruction. Read. register 1 Read. The final path PC 4 Add Reg Shift left 2 Add PCSrc Instrction [3-] Instrction I [25-2] I [2-6] I [5 - ] register register 2 register 2 Registers ALU Zero Reslt ALUOp em Data emtor RegDst ALUSrc em I [5

More information

Millimeter-Wave Multi-Hop Wireless Backhauling for 5G Cellular Networks

Millimeter-Wave Multi-Hop Wireless Backhauling for 5G Cellular Networks 2017 IEEE 85th Vehiclar Technology Conference (VTC-Spring) Millimeter-Wave Mlti-Hop Wireless Backhaling for 5G Celllar Networks B. P. S. Sahoo, Chn-Han Yao, and Hng-Y Wei Gradate Institte of Electrical

More information

IMPLEMENTATION OF OBJECT ORIENTED APPROACH TO MODIFIED ANT ALGORITHM FOR TASK SCHEDULING IN GRID COMPUTING

IMPLEMENTATION OF OBJECT ORIENTED APPROACH TO MODIFIED ANT ALGORITHM FOR TASK SCHEDULING IN GRID COMPUTING International Jornal of Modern Engineering Research (IJMER) www.imer.com Vol.1, Isse1, pp-134-139 ISSN: 2249-6645 IMPLEMENTATION OF OBJECT ORIENTED APPROACH TO MODIFIED ANT ALGORITHM FOR TASK SCHEDULING

More information

Lecture 4: Routing. CSE 222A: Computer Communication Networks Alex C. Snoeren. Thanks: Amin Vahdat

Lecture 4: Routing. CSE 222A: Computer Communication Networks Alex C. Snoeren. Thanks: Amin Vahdat Lectre 4: Roting CSE 222A: Compter Commnication Networks Alex C. Snoeren Thanks: Amin Vahdat Lectre 4 Overview Pop qiz Paxon 95 discssion Brief intro to overlay and active networking 2 End-to-End Roting

More information

Constructing Multiple Light Multicast Trees in WDM Optical Networks

Constructing Multiple Light Multicast Trees in WDM Optical Networks Constrcting Mltiple Light Mlticast Trees in WDM Optical Networks Weifa Liang Department of Compter Science Astralian National University Canberra ACT 0200 Astralia wliang@csaneda Abstract Mlticast roting

More information

Efficient Scheduling for Periodic Aggregation Queries in Multihop Sensor Networks

Efficient Scheduling for Periodic Aggregation Queries in Multihop Sensor Networks 1 Efficient Schedling for Periodic Aggregation Qeries in Mltihop Sensor Networks XiaoHa X, Shaojie Tang, Member, IEEE, XiangYang Li, Senior Member, IEEE Abstract In this work, we stdy periodic qery schedling

More information

5 Performance Evaluation

5 Performance Evaluation 5 Performance Evalation his chapter evalates the performance of the compared to the MIP, and FMIP individal performances. We stdy the packet loss and the latency to restore the downstream and pstream of

More information

The single-cycle design from last time

The single-cycle design from last time lticycle path Last time we saw a single-cycle path and control nit for or simple IPS-based instrction set. A mlticycle processor fies some shortcomings in the single-cycle CPU. Faster instrctions are not

More information

Addressing in Future Internet: Problems, Issues, and Approaches

Addressing in Future Internet: Problems, Issues, and Approaches Addressing in Ftre Internet: Problems, Isses, and Approaches Mltimedia and Mobile commnications Laboratory Seol National University Jaeyong Choi, Chlhyn Park, Hakyng Jng, Taekyong Kwon, Yanghee Choi 19

More information

TAKING THE PULSE OF ICT IN HEALTHCARE

TAKING THE PULSE OF ICT IN HEALTHCARE ICT TODAY THE OFFICIAL TRADE JOURNAL OF BICSI Janary/Febrary 2016 Volme 37, Nmber 1 TAKING THE PULSE OF ICT IN HEALTHCARE + PLUS + High-Power PoE + Using HDBaseT in AV Design for Schools + Focs on Wireless

More information

Date: December 5, 1999 Dist'n: T1E1.4

Date: December 5, 1999 Dist'n: T1E1.4 12/4/99 1 T1E14/99-559 Project: T1E14: VDSL Title: Vectored VDSL (99-559) Contact: J Cioffi, G Ginis, W Y Dept of EE, Stanford U, Stanford, CA 945 Cioffi@stanforded, 1-65-723-215, F: 1-65-724-3652 Date:

More information

DVR 630/650 Series. Video DVR 630/650 Series. 8/16-Channel real-time recording with CIF resolution. Flexible viewing with two monitor outputs

DVR 630/650 Series. Video DVR 630/650 Series. 8/16-Channel real-time recording with CIF resolution. Flexible viewing with two monitor outputs Video DVR 630/650 Series DVR 630/650 Series 8/16-Channel real-time recording with resoltion Flexible viewing with two monitor otpts Remote viewing, playback, control, and configration Easy Pan/Tilt/Zoom

More information

Tdb: A Source-level Debugger for Dynamically Translated Programs

Tdb: A Source-level Debugger for Dynamically Translated Programs Tdb: A Sorce-level Debgger for Dynamically Translated Programs Naveen Kmar, Brce R. Childers, and Mary Lo Soffa Department of Compter Science University of Pittsbrgh Pittsbrgh, Pennsylvania 15260 {naveen,

More information

Workshop. Improving the Bus Network

Workshop. Improving the Bus Network Workshop Improving the Bs Network Workshop Objectives To consider a range of objectives and proposals that cold improve the bs system in Brisbane to contribte to pblic (inclding RACQ members ) mobility

More information

Lecture 13: Traffic Engineering

Lecture 13: Traffic Engineering Lectre 13: Traffic Engineering CSE 222A: Compter Commnication Networks Alex C. Snoeren Thanks: Mike Freedman, Nick Feamster, and Ming Zhang Lectre 13 Overview Dealing with mltiple paths Mltihoming Entact

More information

CS 153 Design of Operating Systems Spring 18

CS 153 Design of Operating Systems Spring 18 CS 153 Design of Operating Systems Spring 18 Lectre 2: Historical Perspective Instrctor: Chengy Song Slide contribtions from Nael Ab-Ghazaleh, Harsha Madhyvasta and Zhiyn Qian Last time What is an OS?

More information

Secure Biometric-Based Authentication for Cloud Computing

Secure Biometric-Based Authentication for Cloud Computing Secre Biometric-Based Athentication for Clod Compting Kok-Seng Wong * and Myng Ho Kim School of Compter Science and Engineering, Soongsil University, Sangdo-Dong Dongjak-G, 156-743 Seol Korea {kswong,kmh}@ss.ac.kr

More information

Optimal Sampling in Compressed Sensing

Optimal Sampling in Compressed Sensing Optimal Sampling in Compressed Sensing Joyita Dtta Introdction Compressed sensing allows s to recover objects reasonably well from highly ndersampled data, in spite of violating the Nyqist criterion. In

More information

webinar series

webinar series Ethernet@Atomotive webinar series Moving Forward: Tool Spported Development for Atomotive Ethernet in Time Sensitive Networks V1.06 2016-07-04 Agenda Introdction 3 Recap: Physical layers, network topology

More information

Congestion-adaptive Data Collection with Accuracy Guarantee in Cyber-Physical Systems

Congestion-adaptive Data Collection with Accuracy Guarantee in Cyber-Physical Systems Congestion-adaptive Data Collection with Accracy Garantee in Cyber-Physical Systems Nematollah Iri, Lei Y, Haiying Shen, Gregori Calfield Department of Electrical and Compter Engineering, Clemson University,

More information

Features. ICMS Integrated Corrosion Management System

Features. ICMS Integrated Corrosion Management System ICMS Integrated Corrosion System Featres Total Corrosion Data Data Exhange with DCS/PCS/SCADA Systems Correlate Corrosion & Process Data Enables Highly Cost-Effective Asset Designed Specifically for Corrosion

More information

dss-ip Manual digitalstrom Server-IP Operation & Settings

dss-ip Manual digitalstrom Server-IP Operation & Settings dss-ip digitalstrom Server-IP Manal Operation & Settings Table of Contents digitalstrom Table of Contents 1 Fnction and Intended Use... 3 1.1 Setting p, Calling p and Operating... 3 1.2 Reqirements...

More information

Tu P7 15 First-arrival Traveltime Tomography with Modified Total Variation Regularization

Tu P7 15 First-arrival Traveltime Tomography with Modified Total Variation Regularization T P7 15 First-arrival Traveltime Tomography with Modified Total Variation Reglarization W. Jiang* (University of Science and Technology of China) & J. Zhang (University of Science and Technology of China)

More information

Bias of Higher Order Predictive Interpolation for Sub-pixel Registration

Bias of Higher Order Predictive Interpolation for Sub-pixel Registration Bias of Higher Order Predictive Interpolation for Sb-pixel Registration Donald G Bailey Institte of Information Sciences and Technology Massey University Palmerston North, New Zealand D.G.Bailey@massey.ac.nz

More information

Broadcasting XORs: On the Application of Network Coding in Access Point-to-Multipoint Networks

Broadcasting XORs: On the Application of Network Coding in Access Point-to-Multipoint Networks Broadcasting XORs: On the Application of Network Coding in Access Point-to-Mltipoint Networks The MIT Faclty has made this article openly available Please share how this access benefits yo Yor story matters

More information

ICMS3 Integrated Corrosion Management System

ICMS3 Integrated Corrosion Management System Integrated System Featres Total Data Data Exhange with DCS/PCS/SCADA Systems Correlate & Process Data Enables Highly Cost-Effective Asset Designed Specifically for Personnel Fll Client- Operation The Integrated

More information

Chapter 7 TOPOLOGY CONTROL

Chapter 7 TOPOLOGY CONTROL Chapter TOPOLOGY CONTROL Oeriew Topology Control Gabriel Graph et al. XTC Interference SINR & Schedling Complexity Distribted Compting Grop Mobile Compting Winter 00 / 00 Distribted Compting Grop MOBILE

More information

QoS-driven Runtime Adaptation of Service Oriented Architectures

QoS-driven Runtime Adaptation of Service Oriented Architectures Qo-driven Rntime Adaptation of ervice Oriented Architectres Valeria ardellini 1 Emiliano asalicchio 1 Vincenzo Grassi 1 Francesco Lo Presti 1 Raffaela Mirandola 2 1 Dipartimento di Informatica, istemi

More information

Pavlin and Daniel D. Corkill. Department of Computer and Information Science University of Massachusetts Amherst, Massachusetts 01003

Pavlin and Daniel D. Corkill. Department of Computer and Information Science University of Massachusetts Amherst, Massachusetts 01003 From: AAAI-84 Proceedings. Copyright 1984, AAAI (www.aaai.org). All rights reserved. SELECTIVE ABSTRACTION OF AI SYSTEM ACTIVITY Jasmina Pavlin and Daniel D. Corkill Department of Compter and Information

More information

A choice relation framework for supporting category-partition test case generation

A choice relation framework for supporting category-partition test case generation Title A choice relation framework for spporting category-partition test case generation Athor(s) Chen, TY; Poon, PL; Tse, TH Citation Ieee Transactions On Software Engineering, 2003, v. 29 n. 7, p. 577-593

More information

The Impact of Avatar Mobility on Distributed Server Assignment for Delivering Mobile Immersive Communication Environment

The Impact of Avatar Mobility on Distributed Server Assignment for Delivering Mobile Immersive Communication Environment This fll text paper was peer reviewed at the direction of IEEE Commnications Society sbject matter experts for pblication in the ICC 27 proceedings. The Impact of Avatar Mobility on Distribted Server Assignment

More information

IP Multicast Fault Recovery in PIM over OSPF

IP Multicast Fault Recovery in PIM over OSPF 1 IP Mlticast Falt Recovery in PIM over OSPF Abstract Relatively little attention has been given to nderstanding the falt recovery characteristics and performance tning of native IP mlticast networks.

More information

Resolving Linkage Anomalies in Extracted Software System Models

Resolving Linkage Anomalies in Extracted Software System Models Resolving Linkage Anomalies in Extracted Software System Models Jingwei W and Richard C. Holt School of Compter Science University of Waterloo Waterloo, Canada j25w, holt @plg.waterloo.ca Abstract Program

More information

Making Full Use of Multi-Core ECUs with AUTOSAR Basic Software Distribution

Making Full Use of Multi-Core ECUs with AUTOSAR Basic Software Distribution Making Fll Use of Mlti-Core ECUs with AUTOSAR Basic Software Distribtion Webinar V0.1 2018-09-07 Agenda Motivation for Mlti-Core AUTOSAR Standard: SWC-Split MICROSAR Extension: BSW-Split BSW-Split: Technical

More information

LWIP and Wi-Fi Boost Flow Control

LWIP and Wi-Fi Boost Flow Control LWIP and Wi-Fi Boost Flow Control David López-Pérez 1, Jonathan Ling 1, Bong Ho Kim 1, Vasdevan Sbramanian 1, Satish Kangovi 1, Ming Ding 2 1 Nokia Bell Laboratories 2 Data61, Astralia Abstract 3GPP LWIP

More information

The extra single-cycle adders

The extra single-cycle adders lticycle Datapath As an added bons, we can eliminate some of the etra hardware from the single-cycle path. We will restrict orselves to sing each fnctional nit once per cycle, jst like before. Bt since

More information

Lecture 10. Diffraction. incident

Lecture 10. Diffraction. incident 1 Introdction Lectre 1 Diffraction It is qite often the case that no line-of-sight path exists between a cell phone and a basestation. In other words there are no basestations that the cstomer can see

More information

Standard. 8029HEPTA DataCenter. Because every fraction of a second counts. network synchronization requiring minimum space. hopf Elektronik GmbH

Standard. 8029HEPTA DataCenter. Because every fraction of a second counts. network synchronization requiring minimum space. hopf Elektronik GmbH 8029HEPTA DataCenter Standard Becase every fraction of a second conts network synchronization reqiring minimm space hopf Elektronik GmbH Nottebohmstraße 41 58511 Lüdenscheid Germany Phone: +49 (0)2351

More information

EMC ViPR. User Guide. Version

EMC ViPR. User Guide. Version EMC ViPR Version 1.1.0 User Gide 302-000-481 01 Copyright 2013-2014 EMC Corporation. All rights reserved. Pblished in USA. Pblished Febrary, 2014 EMC believes the information in this pblication is accrate

More information

Evaluating Influence Diagrams

Evaluating Influence Diagrams Evalating Inflence Diagrams Where we ve been and where we re going Mark Crowley Department of Compter Science University of British Colmbia crowley@cs.bc.ca Agst 31, 2004 Abstract In this paper we will

More information

DIRECT AND PROGRESSIVE RECONSTRUCTION OF DUAL PHOTOGRAPHY IMAGES

DIRECT AND PROGRESSIVE RECONSTRUCTION OF DUAL PHOTOGRAPHY IMAGES DIRECT AND PROGRESSIVE RECONSTRUCTION OF DUAL PHOTOGRAPHY IMAGES Binh-Son Ha 1 Imari Sato 2 Kok-Lim Low 1 1 National University of Singapore 2 National Institte of Informatics, Tokyo, Japan ABSTRACT Dal

More information

STABILITY OF SIMULTANEOUS RECURRENT NEURAL NETWORK DYNAMICS FOR STATIC OPTIMIZATION

STABILITY OF SIMULTANEOUS RECURRENT NEURAL NETWORK DYNAMICS FOR STATIC OPTIMIZATION STABILITY OF SIMULTANEOUS RECURRENT NEURAL NETWOR DYNAMICS FOR STATIC OPTIMIZATION Grsel Serpen and Yifeng X Electrical Engineering and Compter Science Department, The University of Toledo, Toledo, OH

More information

Content Safety Precaution... 4 Getting started... 7 Input method... 9 Using the Menus Use of USB Maintenance & Safety...

Content Safety Precaution... 4 Getting started... 7 Input method... 9 Using the Menus Use of USB Maintenance & Safety... STAR -1- Content 1. Safety Precation... 4 2. Getting started... 7 Installing the cards and the Battery... 7 Charging the Battery... 8 3. Inpt method... 9 To Shift Entry Methods... 9 Nmeric and English

More information

Fault Tolerance in Hypercubes

Fault Tolerance in Hypercubes Falt Tolerance in Hypercbes Shobana Balakrishnan, Füsn Özgüner, and Baback A. Izadi Department of Electrical Engineering, The Ohio State University, Colmbs, OH 40, USA Abstract: This paper describes different

More information

Illumina LIMS. Software Guide. For Research Use Only. Not for use in diagnostic procedures. Document # June 2017 ILLUMINA PROPRIETARY

Illumina LIMS. Software Guide. For Research Use Only. Not for use in diagnostic procedures. Document # June 2017 ILLUMINA PROPRIETARY Illmina LIMS Software Gide Jne 2017 ILLUMINA PROPRIETARY This docment and its contents are proprietary to Illmina, Inc. and its affiliates ("Illmina"), and are intended solely for the contractal se of

More information

TRUSTED WIRELESS HEALTH A New Approach to Medical Grade Wireless

TRUSTED WIRELESS HEALTH A New Approach to Medical Grade Wireless By Mitchell Ross TRUSTED WIRELESS HEALTH A New Approach to Medical Grade Wireless Several crrent trends give case to rethink the design of wireless systems in medical bildings. Increasingly, patients are

More information

FLARE: Coordinated Rate Adaptation for HTTP Adaptive Streaming in Cellular Networks

FLARE: Coordinated Rate Adaptation for HTTP Adaptive Streaming in Cellular Networks FLARE: Coordinated Rate Adaptation for HTTP Adaptive Streaming in Celllar Networks Yongbin Im, Jinyong Han, Ji Hoon Lee, Yoon Kwon, Carlee Joe-Wong, Ted Taekyong Kwon, and Sangtae Ha University of Colorado

More information

MultiView: Improving Trust in Group Video Conferencing Through Spatial Faithfulness David T. Nguyen, John F. Canny

MultiView: Improving Trust in Group Video Conferencing Through Spatial Faithfulness David T. Nguyen, John F. Canny MltiView: Improving Trst in Grop Video Conferencing Throgh Spatial Faithflness David T. Ngyen, John F. Canny CHI 2007, April 28 May 3, 2007, San Jose, California, USA Presented by: Stefan Stojanoski 1529445

More information

Hardware-Accelerated Free-Form Deformation

Hardware-Accelerated Free-Form Deformation Hardware-Accelerated Free-Form Deformation Clint Cha and Ulrich Nemann Compter Science Department Integrated Media Systems Center University of Sothern California Abstract Hardware-acceleration for geometric

More information

Content Content Introduction

Content Content Introduction Content Content Introdction...................................................................... 3 Roles in the provisioning process............................................................... 4 Server

More information

EMC AppSync. User Guide. Version REV 01

EMC AppSync. User Guide. Version REV 01 EMC AppSync Version 1.5.0 User Gide 300-999-948 REV 01 Copyright 2012-2013 EMC Corporation. All rights reserved. Pblished in USA. EMC believes the information in this pblication is accrate as of its pblication

More information

Pipelined van Emde Boas Tree: Algorithms, Analysis, and Applications

Pipelined van Emde Boas Tree: Algorithms, Analysis, and Applications This fll text paper was peer reviewed at the direction of IEEE Commnications Society sbject matter experts for pblication in the IEEE INFOCOM 007 proceedings Pipelined van Emde Boas Tree: Algorithms, Analysis,

More information

Functions of Combinational Logic

Functions of Combinational Logic CHPTER 6 Fnctions of Combinational Logic CHPTER OUTLINE 6 6 6 6 6 5 6 6 6 7 6 8 6 9 6 6 Half and Fll dders Parallel inary dders Ripple Carry and Look-head Carry dders Comparators Decoders Encoders Code

More information

Constructing and Comparing User Mobility Profiles for Location-based Services

Constructing and Comparing User Mobility Profiles for Location-based Services Constrcting and Comparing User Mobility Profiles for Location-based Services Xihi Chen Interdisciplinary Centre for Secrity, Reliability and Trst, University of Lxemborg Jn Pang Compter Science and Commnications,

More information

EMPOWERING SCIENTIFIC DISCOVERY BY DISTRIBUTED DATA MINING ON A GRID INFRASTRUCTURE

EMPOWERING SCIENTIFIC DISCOVERY BY DISTRIBUTED DATA MINING ON A GRID INFRASTRUCTURE EMPOWERING SCIENTIFIC DISCOVERY BY DISTRIBUTED DATA MINING ON A GRID INFRASTRUCTURE A PROPOSAL FOR DOCTORAL RESEARCH by Haimonti Dtta SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE

More information

The Volcano Optimizer Generator: Extensibility and Efficient Search

The Volcano Optimizer Generator: Extensibility and Efficient Search The Volcano Optimizer Generator: Extensibility and Efficient Search - Prithvi Lakshminarayanan - 301313262 Athors Goetz Graefe, Portland State University Won the Most Inflential Paper award in 1993 Worked

More information

History Slicing: Assisting Code-Evolution Tasks

History Slicing: Assisting Code-Evolution Tasks History Slicing: Assisting Code-Evoltion Tasks Francisco Servant Department of Informatics University of California, Irvine Irvine, CA, U.S.A. 92697-3440 fservant@ics.ci.ed James A. Jones Department of

More information

Blended Deformable Models

Blended Deformable Models Blended Deformable Models (In IEEE Trans. Pattern Analysis and Machine Intelligence, April 996, 8:4, pp. 443-448) Doglas DeCarlo and Dimitri Metaxas Department of Compter & Information Science University

More information

AN A. GPON Optical Network Terminal. Product Manual. Version: A/1. FiberHome Telecommunication Technologies Co., Ltd.

AN A. GPON Optical Network Terminal. Product Manual. Version: A/1. FiberHome Telecommunication Technologies Co., Ltd. AN5506-01-A GPON Optical Network Terminal Prodct Manal Version: A/1 FiberHome Telecommnication Technologies Co., Ltd. April 2017 Thank yo for choosing or prodcts. We appreciate yor bsiness. Yor satisfaction

More information

Master for Co-Simulation Using FMI

Master for Co-Simulation Using FMI Master for Co-Simlation Using FMI Jens Bastian Christoph Claß Ssann Wolf Peter Schneider Franhofer Institte for Integrated Circits IIS / Design Atomation Division EAS Zenerstraße 38, 69 Dresden, Germany

More information

Availability Analysis of Application Servers Using Software Rejuvenation and Virtualization

Availability Analysis of Application Servers Using Software Rejuvenation and Virtualization Thein T, Park J S. Availability analysis of application servers sing software rejvenation and virtalization. JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY 24(2): 339 346 Mar. 2009 Availability Analysis of

More information

A Recovery Algorithm for Reliable Multicasting in Reliable Networks

A Recovery Algorithm for Reliable Multicasting in Reliable Networks A Recovery Algorithm for Reliable Mlticasting in Reliable Networks Danyang Zhang Sibabrata Ray Dept. of Compter Science University of Alabama Tscaloosa, AL 35487 {dzhang, sib}@cs.a.ed Ragopal Kannan S.

More information

Discrete Cost Multicommodity Network Optimization Problems and Exact Solution Methods

Discrete Cost Multicommodity Network Optimization Problems and Exact Solution Methods Annals of Operations Research 106, 19 46, 2001 2002 Klwer Academic Pblishers. Manfactred in The Netherlands. Discrete Cost Mlticommodity Network Optimization Problems and Exact Soltion Methods MICHEL MINOUX

More information

Comparison of memory write policies for NoC based Multicore Cache Coherent Systems

Comparison of memory write policies for NoC based Multicore Cache Coherent Systems Comparison of memory write policies for NoC based Mlticore Cache Coherent Systems Pierre Gironnet de Massas, Frederic Petrot System-Level Synthesis Grop TIMA Laboratory 46, Av Felix Viallet, 38031 Grenoble,

More information

COMPOSITION OF STABLE SET POLYHEDRA

COMPOSITION OF STABLE SET POLYHEDRA COMPOSITION OF STABLE SET POLYHEDRA Benjamin McClosky and Illya V. Hicks Department of Comptational and Applied Mathematics Rice University November 30, 2007 Abstract Barahona and Mahjob fond a defining

More information

Uncertainty Determination for Dimensional Measurements with Computed Tomography

Uncertainty Determination for Dimensional Measurements with Computed Tomography Uncertainty Determination for Dimensional Measrements with Compted Tomography Kim Kiekens 1,, Tan Ye 1,, Frank Welkenhyzen, Jean-Pierre Krth, Wim Dewlf 1, 1 Grop T even University College, KU even Association

More information

A GENERIC MODEL OF A BASE-ISOLATED BUILDING

A GENERIC MODEL OF A BASE-ISOLATED BUILDING Chapter 5 A GENERIC MODEL OF A BASE-ISOLATED BUILDING This chapter draws together the work o Chapters 3 and 4 and describes the assembly o a generic model o a base-isolated bilding. The irst section describes

More information

VirtuOS: an operating system with kernel virtualization

VirtuOS: an operating system with kernel virtualization VirtOS: an operating system with kernel virtalization Rslan Nikolaev, Godmar Back SOSP '13 Proceedings of the Twenty-Forth ACM Symposim on Oper ating Systems Principles 이영석, 신현호, 박재완 Index Motivation Design

More information

Appearance Based Tracking with Background Subtraction

Appearance Based Tracking with Background Subtraction The 8th International Conference on Compter Science & Edcation (ICCSE 213) April 26-28, 213. Colombo, Sri Lanka SD1.4 Appearance Based Tracking with Backgrond Sbtraction Dileepa Joseph Jayamanne Electronic

More information

Multiple-Choice Test Chapter Golden Section Search Method Optimization COMPLETE SOLUTION SET

Multiple-Choice Test Chapter Golden Section Search Method Optimization COMPLETE SOLUTION SET Mltiple-Choice Test Chapter 09.0 Golden Section Search Method Optimization COMPLETE SOLUTION SET. Which o the ollowing statements is incorrect regarding the Eqal Interval Search and Golden Section Search

More information

CAPL Scripting Quickstart

CAPL Scripting Quickstart CAPL Scripting Qickstart CAPL (Commnication Access Programming Langage) For CANalyzer and CANoe V1.01 2015-12-03 Agenda Important information before getting started 3 Visal Seqencer (GUI based programming

More information

Internet Network Management and Policy Recommendations

Internet Network Management and Policy Recommendations Internet Network Management and Policy Recommendations Asia-Pacific Telecommnity (APT) UNESCAP Sb-Regional Workshop on Internet Traffic Management for the Asia- Pacific Information Sperhighway, 7-8 December

More information

An Introduction to GPU Computing. Aaron Coutino MFCF

An Introduction to GPU Computing. Aaron Coutino MFCF An Introdction to GPU Compting Aaron Cotino acotino@waterloo.ca MFCF What is a GPU? A GPU (Graphical Processing Unit) is a special type of processor that was designed to render and maniplate textres. They

More information

CS 153 Design of Operating Systems Spring 18

CS 153 Design of Operating Systems Spring 18 CS 153 Design of Operating Systems Spring 18 Lectre 12: Deadlock Instrctor: Chengy Song Slide contribtions from Nael Ab-Ghazaleh, Harsha Madhyvasta and Zhiyn Qian Deadlock the deadly embrace! Synchronization

More information

AUTOMATIC REGISTRATION FOR REPEAT-TRACK INSAR DATA PROCESSING

AUTOMATIC REGISTRATION FOR REPEAT-TRACK INSAR DATA PROCESSING AUTOMATIC REGISTRATION FOR REPEAT-TRACK INSAR DATA PROCESSING Mingsheng LIAO, Li ZHANG, Zxn ZHANG, Jiangqing ZHANG Whan Technical University of Srveying and Mapping, Natinal Lab. for Information Eng. in

More information

CS 153 Design of Operating Systems Spring 18

CS 153 Design of Operating Systems Spring 18 CS 53 Design of Operating Systems Spring 8 Lectre 2: Virtal Memory Instrctor: Chengy Song Slide contribtions from Nael Ab-Ghazaleh, Harsha Madhyvasta and Zhiyn Qian Recap: cache Well-written programs exhibit

More information

Computer-Aided Mechanical Design Using Configuration Spaces

Computer-Aided Mechanical Design Using Configuration Spaces Compter-Aided Mechanical Design Using Configration Spaces Leo Joskowicz Institte of Compter Science The Hebrew University Jersalem 91904, Israel E-mail: josko@cs.hji.ac.il Elisha Sacks (corresponding athor)

More information

IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 6, NO. 5, MAY On the Analysis of the Bluetooth Time Division Duplex Mechanism

IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 6, NO. 5, MAY On the Analysis of the Bluetooth Time Division Duplex Mechanism IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 6, NO. 5, MAY 2007 1 On the Analysis of the Bletooth Time Division Dplex Mechanism Gil Zssman Member, IEEE, Adrian Segall Fellow, IEEE, and Uri Yechiali

More information

Vector Logger Cloud. VECTOR GB Ltd Conference, 28th Sept, 2017 V

Vector Logger Cloud. VECTOR GB Ltd Conference, 28th Sept, 2017 V Vector Logger Clod VECTOR GB Ltd Conference, 28th Sept, 2017 V1.0 2017-09-27 Agenda Challenges Vector Logger Clod Secrity Aspects Data Acqisition Policy Conclsion 2 Vector Logger Clod Challenges Growing

More information

EMC VNX Series. Problem Resolution Roadmap for VNX with ESRS for VNX and Connect Home. Version VNX1, VNX2 P/N REV. 03

EMC VNX Series. Problem Resolution Roadmap for VNX with ESRS for VNX and Connect Home. Version VNX1, VNX2 P/N REV. 03 EMC VNX Series Version VNX1, VNX2 Problem Resoltion Roadmap for VNX with ESRS for VNX and Connect Home P/N 300-014-335 REV. 03 Copyright 2012-2014 EMC Corporation. All rights reserved. Pblished in USA.

More information

Local Run Manager. Software Reference Guide for MiSeqDx

Local Run Manager. Software Reference Guide for MiSeqDx Local Rn Manager Software Reference Gide for MiSeqDx Local Rn Manager Overview 3 Dashboard Overview 4 Administrative Settings and Tasks 7 Workflow Overview 12 Technical Assistance 17 Docment # 1000000011880

More information

Image Enhancement in the Frequency Domain Periodicity and the need for Padding:

Image Enhancement in the Frequency Domain Periodicity and the need for Padding: Prepared B: Dr. Hasan Demirel PhD Image Enhancement in the Freqenc Domain Periodicit and the need for Padding: Periodicit propert of the DFT: The discrete Forier Transform and the inverse Forier transforms

More information