A Northbound Interface for Power Management in Next Generation Network Devices
|
|
- Brian Goodman
- 6 years ago
- Views:
Transcription
1 A Northbound Interface for Power Management in Next Generation Network Devices Raffaele Bolla 1, Roberto Bruschi 1, Franco Davoli 1, Pasquale Donadio 2, Leonardo Fialho 3, Martin Collier 3, Alfio Lombardo 4, Diego Reforgiato 4, Vincenzo Riccobene 4, Tivadar Szemethy CNIT, Genoa, Italy 2- Alcatel-Lucent, Italy 3- Dublin City University, Dublin, Ireland 4- Lightcomm, Catania, Italy 5- Netvisor, Budapest, Hungary Abstract Recently, a number of approaches based on dynamic power management techniques have been proposed to reduce the energy consumption of telecommunication networks and devices. They are able to optimize the tradeoff between network performance and energy requirements. It is possible to execute and extend these techniques to the whole network, by using local control policies together with energy-aware routing and traffic engineering. However, the lack of a standardized representation of the energy-aware capabilities of heterogeneous networking equipment makes their deployment confusing and impractical. To this aim, we have proposed a novel framework, the Green Abstraction Layer (), whose purpose is to define a multi-layered abstraction interface for the hardware and physical resources, where energy management actions are directly performed. Therefore, the syntax can be exposed to the platform-independent logical representation commonly used in network control protocols. Given the internal architectural complexity and heterogeneity of many network devices, the approach is based on a hierarchical decomposition, where each level provides an abstract and aggregated representation of internal components. 1 Introduction Network energy efficiency has become an aspect of paramount importance for Telco operators and Internet Service Providers (ISPs) across all network segments (wired and wireless), owing to the increasing power consumption exhibited over the years [1] and to the increasing cost of energy. Estimates of public organizations [2] provide alarming figures in the Business-As-Usual case; for the European Telecommunication companies only, the estimated energy requirement in 2020 is 35.8 TWh, which represents an increase of over 67% in a decade. The increasing user traffic and router capacities are two of the main reasons for such high power consumption. They are not compensated for by a corresponding increase in silicon energy efficiency. Moreover, the vast majority of currently deployed network links and devices are designed to operate (and, consequently, to consume energy) constantly at their maximum capacity, even though their average utilization lies far below the maximum [1,3]. These observations have motivated the choice of adapting The work described in this paper was performed with the support of the ECONET project (low Energy Consumption NETworks), funded by the EU in FP7 call 5 (grant agreement no ). Homepage:
2 network energy requirements to the actual traffic profiles [4,5]. This can be realized by providing Power Management Primitives (PMPs) into the hardware of networking devices, where energy absorption physically happens. PMPs allow aggressively reducing the energy consumption of networking devices (or some parts of them) by putting them into standby states when not in use, or by decreasing their maximum performance in the presence of low traffic volumes. The reader may notice that, on the one hand, it is obvious that the maximal power saving is obtained when the equipment is turned off; however, under such a condition the performance is actually zero, as well. On the other hand, it is also clear that the best performance is obtained when the device operates with no power constraint. Therefore, power-managed devices need control loops able to dynamically tune hardware capabilities to provide the required Quality of Service (QoS) level to the incoming traffic with the minimal power consumption. It is also worth noting that PMPs are features locally available in network nodes, and their efficiency may heavily depend on the specific implementation and low-level details of the hardware platforms of devices. Actually, providing each network device with its own independent control loop or Local Control Policy () would appear straightforward. According to the specific features of the local device, the control loop may dynamically orchestrate the configuration of internal components (e.g., line-cards, network processors, etc.) to meet the desired QoS with the minimum consumption. However, should each device independently perform energy optimizations, then the overall network consumption might result much higher than in the case of cooperation among nodes, the satisfaction of end-to-end network QoS requirements might not be guaranteed. Along this direction, several approaches [6-8] have been recently proposed in order to extend current routing, traffic engineering and network management policies beyond classical network QoS metrics, and to explicitly consider also the energy consumption of the whole network. For instance, moving traffic flows towards alternative network paths may either reduce the load of some nodes (or part of them), by putting them into standby mode [6-7], or trigger Power Scaling (PS) primitives in a more effective way. Typical examples of these kinds of processes that can be extended to deal with green metrics are management protocols (e.g., TMF in metropolitan networks, or OpenFlow in datacenters), or traffic engineering and routing algorithms/protocols (e.g., OSPF-TE/RSVP-TE). We will refer to these approaches as Network-wide Control Policies (NCP). Even if potentially much more effective than s, NCPs suffer from some drawbacks. First, the NCPs can exhibit much higher feedback/convergence delays. Secondly, a further problem in interfacing NCPs and the power management of physical entities lies in the fact that network control protocols generally work on top of logical resources, often losing the details on how they can be mapped on devices internal hardware, where power consumption and management physically happen. Thus, NCPs may drive power configurations of network devices only by means of such logical resources, but they would need to know the mapping between logical resources and internal device components, in order to understand the effects of their action. Finally, NCPs often represent a network device simply as a node in a graph, whose edges are the (virtual/physical) network links. Such simplistic representation does not allow maintaining the knowledge of some hardware peculiarities that may be crucial for reducing consumption. For example, let us consider a device with some line-cards, each one with multiple ports: putting all the links of the
3 same line-card to sleep is much more profitable than other configurations, since an entire line-card can enter sleep mode. All these points suggest the possibility of jointly adopting s and NCPs in order to optimize the device energy consumption in a hierarchical way. It is easy to conclude that there is still a significant gap between hardware power management and NCPs, as well as some open issues on which control loops should be adopted, and on how to effectively manage the relationships among multiple (local and/or network-wide) control loops. Moreover, an abstract representation is needed, independently of the details of the specific manufacturers implementations. With all this in mind, the main contribution of this paper consists in the specification and initial experimental characterization of a layer that provides a way to expose green networking capabilities of devices toward the network control plane. Such a layer consists of a novel standard interface, the Green Abstraction Layer () [9], conceived by the ECONET consortium ( The specific goal of the is not to directly provide additional energy gains to energy-aware devices, but rather to offer a simple and well-defined substrate that presents the following peculiar aspects: (i) (ii) (iii) a common and simple way for representing power management capabilities available in heterogeneous data plane hardware, a framework for mapping the capabilities of power-managed physical entities at the data-plane with the logical resources used by control processes, a reference control chain allowing a consistent hierarchical organization of multiple local and network-wide control loops. As such, the target of this paper is to define a complete control interface that will allow the management of the trade-off between power and network performance at the system level. In other words, the is meant to make control loops able to acquire information on which power management settings are available at the data-plane and on their potential effect in terms of consumption and network performance metrics, in order to choose the most suitable configuration. Similarly to other standards in the same field, like the IETF Energy MANagement (EMAN these types of information are formalized through the concept of energy-aware states. However, the power state description and characterization that existing standards provide is clearly insufficient for any remote control decisions (e.g., lack of network-specific performance tradeoff indications in power state descriptions). The remainder of this paper is organized as follows: Section 2 describes the general architecture and structure of the. Section 3 introduces the representation of power management primitives. Section 4 defines energy-aware states and their relevant parameters. Section 5 ends the paper with conclusions and future research directions where we are heading. 2 The architecture In order to map the energy consumption of physical components with the performance and the functionalities at the network logical level, the is provided with a hierarchical layered view of devices internal organization, which represents the components at the various levels as a tree of entities (Figure 1.a). Since this kind of representation is already largely adopted in the forwarding and configuration information databases of network devices, the does not explicitly define various aspects (e.g., the entity identifiers, the maximum depth of the tree, etc.), in order to favor its integration with these existing device-internal databases.
4 The root entity populates layer 1 and corresponds to the entire node. Layer 0 includes multiple entities representing the device logical resources, which are functional objects/instances (e.g., IP links, tunnel/vlan virtual interfaces, etc.) commonly used for representing local protocol configurations, and exposing them toward control and management processes. Since the nature, the number, and the relationship among logical resources may vary a lot depending on the network protocols implemented, layer 0 was designed to handle these resources as a simple flat list. The role of layer 0 is then simply representing logical resources to the Control Plane in the syntax, as a means to allow the mediation between Control Plane processes and hardware specific characteristics. The mapping of logical resources representing per-resource management logic, the managed objects and their s is specific to the each given device architecture. For instance, as shown in Figure 1.b, the logical resources in a router correspond to the IP interfaces configured on both physical and virtual (e.g., VLAN) ports. The presence of layers 0 and 1 is mandatory; layers 2,, n are optional, and are meant to represent the internal hardware organization of the device at various levels of aggregation. For instance, as shown in Fig. 1.b, the hardware architecture of the router prototype to be described in Section 5 can be decomposed in 2 further layers namely, the and Hardware Components. A similar breakdown is also adopted in reference [9] (which also contains the definition of the methods). In more details, the may even represent as entities only those hardware components/subsystems able to trade off their power consumption and their performance, available functionalities, and responsiveness levels. In such a case, every node of the tree corresponds to a. As power management primitives reside at the lowest levels of the hierarchy, every leaf of the tree connects with a Convergence Layer, which includes all the needed means (e.g., drivers/agents, etc.) to interface the with PMPs of hardware objects. Leaf nodes may reside at different layers of the hierarchy for two main reasons: some hardware elements may need to be accessible from higher hierarchical levels (e.g., fans in a chassis); some manufacturers might prefer not to expose the internal organization and the subcomponents of a composite part of the device. Starting from the n-th layer, the is hierarchically applied in order to give the possibility to control policies at the i-th layer to orchestrate the configurations of the underlying entities in order to aggregate their operating behaviors, and to expose them as a single energy-aware entity at the upper layer. It is worth noting that the role of the interface does not depend on the hierarchical level where it is applied, but it simply allows acquiring information from or setting a configuration to an entity with PMPs.
5 Layer 3: HW Components Layer 2: Layer 1: Device Layer 0: Logical Resources Network Control and Management Processes (e.g., Signalling Protocols) Layer 0: Logical Resources Layer 1: Device Root Layer 2: 1 st level HW Res. Layer 3: 2 nd level HW Res a Entity 1.1 DROP data-plane FE1 CLI/Routing Protocols eth1 eth2 eth40 Entity 1.2 FE2 Entity 1 Entity 1.3 FE3 Layer n: (n-1) th level HW Res m Convergence Layer 1...m (a) Figure 1: Abstract view of the hierarchical architecture (a), and its instantiation when applied to the prototype introduced in section 5 (b). This process terminates at layer 1, where the highest handles the high-level configuration of the device, and needs to expose a simplified view of it, by mapping and decomposing the impact of possible power settings to the device logical resources residing at layer 0, which are meant to provide per-resource management logic through the syntax. For instance, with reference to Figure 1.b, logical resources of the router prototype correspond to the available IP (virtual/physical) ports. The role of the layer-1 in this case is to map/decompose the power management configurations of layer-2 elements on a per IP-port basis at level 0, with criteria that obviously depend on the specific device architecture. The possibility of referring to logical resources to acquire and set information regarding energy configurations is a key aspect of the entire architecture, since it allows: - Accessing and/or modifying the device energy-aware configuration on the part of NCPs or even a simple command line interface, by working on the same logical/functional objects, which are listed in the forwarding and configuration databases; - Hiding how a functional/logical resource of the device is implemented in the hardware, and which subsystems it involves or shares with other resources. It is also worth noting that the was specifically designed to be agnostic according to the nature of NCP processes and related signalling protocols, since they may vary significantly with respect to the specific network technologies and deployment scenarios. For instance, the does not specify/impose which kind of information needs to be handled by the management plane or by the network control. Rather, the main goal of the is to provide a common substrate for such control and management processes and signalling protocols, which may make the mapping of hardware PMPs to the logical network level easier and uniform across heterogeneous network technologies and architectures. For this (b)
6 reason, the presence of the does not provide any additional energy saving to network devices, but can foster a complete and explicit support to power management operations in network infrastructures. 3 Power Management Primitives PMPs allow dynamically trading off energy consumption and processing performance of hardware components, and they can be classified on the basis of their effect on the working behavior of the managed component. According to [5] and to Figure 2.a, a basic classification is the following: Standby primitives, which consist of PMPs that allow devices or parts of them entering very low energy states, where most of their functionalities are frozen. PS primitives, which allow dynamically reducing the working rate of the component. This is usually accomplished by adopting two basic techniques: Adaptive Rate (AR) and Low Power Idle (LPI). The former deals with the dynamic modulation of the processing capacity of the component. The latter forces links or processing engines to enter low power states when not sending/processing packets, and to quickly switch to a higher power state when sending one or more packets. LPI and AR can be adopted jointly, in order to adapt components performance to current workload requirements, and have different impacts on packet forwarding performance and QoS. As shown in Figure 2.b, AR obviously causes the stretching of packet service times, while the adoption of LPI only introduces an additional delay in packet service, due to the wake-up times. Although PMPs are all built upon the basic hardware techniques cited above, they can be exposed with different operational behaviours. At the generic i-th level of the hierarchy, an entity may provide the management of some aspects/settings to the above /NCP (see Figure 2.b); some other aspects/settings may be operated by the entity itself in an autonomic way (see Figure 2.c). In this case, the entity might already include s internally in lower-level components or directly in the hardware. Starting from these considerations, the operational behaviours of such capabilities can be summarized as follows: Autonomic: the same entity manages (part of) its energy configuration in an autonomic way; Controllable: the modification of the energy configuration (or part of it) requires an explicit action from the upper-layer control process. In this sense, we can classify entities as Autonomic Entities and Controllable Entities (AE and CE). As shown in Figure 2.c, an AE drives its operational behaviour without any explicit actions/configurations from the upper layer control process, which only knows the aggregated characterization of the operational behaviour in terms of energy profile against the traffic offered load. When one or more PMP settings are configurable by the upper layer control process through the, the entity is a CE. Figure 2.b shows the case of a CE, where all PMP settings are explicitly operated by the upper layer, which decides when and how to change the configuration of AR and LPI. Between the two limit operating behaviours in Figures 2.b and 2.c, a number of variants can be identified and traced back to the considerations above. In more detail, an entity may be designed to autonomically throttle between two AR settings (ARi and ARj), which must be configured by controlplane processes. In such a case, each available power profile (e.g., a pair of i and j indexes for the limit AR settings) represents an independent autonomic profile that may be set (and then controlled) by the. Under such assumption the entity has to be considered as a CE.
7 Consumption Consumption Consumption Consumption Consumption Consumption Standby Power Scaling Low Power Idle (LPI) Adaptive Rate (AR) AR 2 AR 1 AR z AR z-1 S k S 1 LPI v LPI 1 (a) Wake-up time Wake-up time Max Rate time (b) AR 1 AR z LPI v AR z AR 2 LPI 1 AR AR 1 zarz-1 AR 2 Offered Load time AR 1 AR z LPI v AR z AR 2 LPI 1 AR AR 1 (c) zarz-1 AR 2 Figure 2: Hardware Power Management Primitives and their trade-off between energy consumption and performance (a), their operating behavior in controlled mode, where the transition among LPI and AR states is explicitly driven by the (b), and in autonomic mode, where no direct interaction with the is needed (c). 4 Energy-Aware States An EAS is an operational power profile mode implemented by the entity that can be configured by upper-layer control processes. According to section 3, under autonomic operations AEs can expose only one EAS describing their aggregate performance/consumption profile, while CEs have two or more available EASs, which may still contain autonomic operations. From a general point of view, all the aspects configurable from the upper-layer control process are translated into EASs, and autonomic behaviours are represented as descriptive parameters of EASs.
8 To assure the correct exchange of information between the entity and control plane processes, the EAS must be represented as a complex data type, which contains information about power consumption, performance, available functionalities, and responsiveness of the entity when working in such a configuration. This information is obviously necessary to s and NCPs in order to decide which EAS provides the desired trade-off between power consumption and network performance. Owing to the adaptive and autonomic nature of some PMP realizations of EASs described in section 3, the overall energy saving and the network performance may heavily depend on incoming traffic volumes and statistical features (traffic burstiness levels, etc.). For this reason, it would be quite hard to precisely know a-priori the overall entity performance when working in an EAS. Therefore, such characterization indexes must be expressed in an objective (and easily measurable) form that may suit different incoming traffic patterns, and may help binding the maximum energy consumption and/or the worst possible level of network performance. The most intuitive solution is to make such indexes describing the high-level effects introduced by power management techniques (i.e., speed reduction, wake-up times, etc.), which do not depend on the system dynamics. We modelled the EAS as a couple of energy-aware Primitive sub-states (PsSs) related to the configuration of Standby and Power Scaling mechanisms: where EAS jk={p j, S k} with 0 j J and 0 k K Sk represents the k-th PsS related to the standby techniques (S-PsS); S0 is the active state, in Sk, with k>0, the entity is sleeping, in SK the device is completely off. Pj represents the j-th PsS related to the Power Scaling techniques; P0 is the PsS for the highest device network performance (it also consumes the highest power), PJ is the PsS with the lowest performance level for the device while still being active (it consumes the least power for the device in active state). See Figure 4 for the definition of PsSs composing EASs. Given the exclusive use of power scaling and standby capabilities, the j-th Power Scaling Primitive sub-states (P-PsS) can be adopted only when the entity is active (S0). When the entity is in standby modes (Sk with k>0), Pj only represents the power scaling configuration that will be used after wakeup. Figure 3: Organization of the Power Scaling and Standby PsSs composing EASs. PsSs are represented by means of a set of parameters that address the following aspects (see Table I): the entity power consumption when working in such a PsS, its performance level and which other EAS
9 (and how) the entity can move to. While the number of attributes needed for S-PsSs is relatively low, and their meaning is quite intuitive, a more complex representation of P-PsSs is required in order to take into account both AR and LPI primitives and their operational behavior. In detail, P-PsSs attributes have been divided into five main categories, as follows. General attributes are meant to provide a snapshot on the considered P-PsS: some performance and maximum power consumption bounds are defined to summarize the aggregated behavior of the entity. Then, more precise data on non-autonomic AR and LPI primitives are given in the AR and LPI attribute categories. The fourth category, related to autonomic attributes, aims at describing the performance of the autonomic energy-aware behavior of the entity. This is done by describing the performance and the consumption of the entity for a number of offered load levels. Finally, state transition attributes give indications on which other P-PsSs the entity can move to from the current one, and the related possible performance decays. S-PsS attributes. Parameter Description Table I: EAS attributes. k Standby Power Saving Maximum Wake-Up time Maximum Sleeping time Wakeup Triggers Sleeping Timer P-PsS attributes. Parameter Identifier number of the current S-PsS. The power gain for being at S k with respect to the maximum power draw of the entity. The maximum time that is needed to wake-up from S k to S 0. The maximum time that is needed to go into S-PsS S k from active S 0. Set of flags that represents which triggers can be used to wake-up the entity. Bit 0: ; Bit 1: wake-onpacket; Bit 2: WoL. Other flags will be defined. Time of inactivity before the entity enters into S k in an autonomic way. Description General attr. AR Specific attr. LPI Specific attr. Autonomic attr. State Trans. attr. j Minimum Gain Maximum Throughput Maximum Throughput LPI Power Gain Power Packet LPI Wakeup Time LPI Sleeping Time LPI Power Autonomic Steps Autonomic Profile Curves Bit Transition PS PS Autonomic PS Service Interruption P-PsS Times Transition P-PsS Transition Service Interruption Times Identifier number of the current Power Scaling PsS. The minimum power gain in P j. with respect to the maximum power draw of the entity. This value is usually reached when the entity works at the maximum throughput in P j. The maximum throughput provided by the entity when working in P j. This value is expressed in packets per second. The maximum throughput provided by the entity when working in P j. This value is expressed in bits per second. The average power gain when the entity is idle in P j with respect to the maximum power draw of the entity. The maximum time that is needed to wake-up from the LPI mode upon packet arrival. The maximum time that is needed to enter into the LPI mode. The average power needed in P j when moving from idle to active modes and vice versa. Granularity used for describing the power and performance profile in autonomic power scaling primitives. This parameter represents the behavior of autonomic PS solutions. It is an array of depth STEPS (i.e., Throttling steps). Each array cell is composed by three values, indicating the offered load, the maximum consumption and the maximum packet service time at that load. Maximum service interruption time that may be incurred when the autonomic power scaling mechanisms switches the entity capacity. Array containing the maximum time intervals needed to move from P j to other P-PsS. The infinity value is used for indicating that a direct transition is not possible. Array containing the maximum service interruption times that may incur while moving from P j to other Power Scaling PsS. The infinity value is used for indicating that a direct transition is not possible.
10 5 A -enabled prototype In the context of the ECONET project, seven prototypes of heterogeneous -enabled devices (spanning from home gateways to high-end switches and routers) have been developed, and their capabilities successfully demonstrated ( In this section, we report some of the results and architectural details that have been experimented with the DROP router, an open-source software router with a modular architecture based on Linux and commercial off-the-shelf hardware. This selection was mainly driven from the openness of the prototype that may help the reader to deepen later on further aspects on the design architecture of this reference implementation. In any case, it is worth remarking here that the functionalities of the are not conceived to provide further gains in energy or performance (a task of control policies), but rather to simplify the interworking between Control and Data planes, to isolate device-specific characteristics, and to ease the task of designers and system developers. All these goals have been achieved in the ECONET prototypes, of which the DROP router described below is a paradigmatic example. As shown in Figure 4, the data-plane of the DROP router is composed by: Three forwarding elements devoted to perform IP layer operations on incoming packets. These elements were realized with Linux software running on commodity servers and Broadcom XLP 832 evaluation boards. An interconnection element - an Openflow switch balancing the incoming traffic among the forwarding elements. An additional Linux server is used as control element of the DROP router, and runs all the control and management signalling protocols. In our experimental setup only forwarding elements provide PMPs in terms of both AR and LPI at network adapters and at processors, where packets are elaborated. elements based on commodity servers also provide standby capabilities. As shown in Figure 1.b, the was realized with a 4 layer-hierarchy: besides the mandatory layers 0 and 1, layer 2 was devoted to manage the forwarding elements, and layer 3 represents the processors and the adapters in the elements. Load Balancing IP routing Demultiplexing Interconnection (RX side) Interconnection (TX side) Figure 4: The DROP architecture.
11 Power Consumption [W] Power Consumption [W] Power Consumption [W] The s at layers 1 and 2 drive the power management of the node, by resulting for each configured IP link in two available EASs with autonomic capabilities. s at layer 2 orchestrate the AR and LPI configurations of both network adapters and processor cores with the strategy developed in [10]. Then, as shown in Figure 5, the at the device level modulates the number of forwarding elements in CLI eth1 eth2 eth40 Entity Autonomic 1 5 Autonomic 2 0 BAU Traffic Load [%] ) Control the number of FEs in standby & move traffic flows among FEs Autonomic Autonomic ) Decompose the energy Standby profiles from hardware 30 BAU entities to logical ones Traffic Load [%] 150 Autonomic 1 Autonomic 2 BAU Traffic Load [%] Entity 1.1 Entity 1.2 Entity 1.3 Interconnection (RX side) Interconnection (TX side) Figure 5: Overview on the role of s in the DROP prototype and of the management of EASs at different hierarchical layers of the. The s at the level of the forwarding elements orchestrate the AR and LPI capabilities to obtain two autonomic P-PsSs and one S-PsS (available only on commoditybased platforms). On its turn, depending on the incoming traffic volume, the at the device level autonomously manages the EASs exposed by lower layer entities and the traffic share sent to each forwarding element, in order to obtain two autonomic P-PsSs, which are then decomposed into a couple of available EASs for each logical link configured on the device. In the autonomic profile graphs, the Business as Usual (BAU) energy consumption has been also reported.
12 standby depending on both the traffic offered load and the EAS configured by the system administrator for each logical resource. 6 Conclusions We have proposed a novel framework, the Green Abstraction Layer, to represent energy-aware capabilities of networking devices. Owing to the heterogeneity of networking equipment by different vendors and to the necessity of deploying energy management policies into existing networks, the definition of a multi-layered abstraction is a key issue to ease the task of manufacturers and network managers, and to close the gap between hardware power management and logical aspects of network control. In this work, we have outlined the definition of such multi-layered abstraction interface from the hardware and physical resources, where energy management actions are directly exerted, to the platformindependent logical representation commonly used in network control protocols. We have described the hierarchical architecture of the, and we have provided the definition of energy-aware states of devices and of their component parts. Finally, we have outlined one of the experimental realizations of the already implemented in the framework of the ECONET project. We are currently further developing the on several network hardware platforms. Moreover, the framework is currently being discussed for standardization within the ETSI Environmental Engineering Technical Committee under Work Item DES/EE References [1] R. Bolla, R. Bruschi, A. Carrega, F. Davoli, D. Suino, C. Vassilakis, A. Zafeiropoulos, Cutting the Energy Bills of Internet Service Providers and Telecoms through Power Management: An Impact Analysis, Computer Networks, vol. 56, no. 10, pp , July [2] Global e-sustainibility Initiative (GeSI), SMART 2020: Enabling the Low Carbon Economy in the Information Age, Smart2020Report.pdf. [3] S. Nedevschi, L. Popa, G. Iannaccone, D. Wetherall, S. Ratnasamy, Reducing Network Energy Consumption via Sleeping and Rate-Adaptation, Proc. 5th USENIX Symp. on Networked Systems Design and Implementation, San Francisco, CA, April 2008, pp [4] R. Bolla, R. Bruschi, K. Christensen., F. Cucchietti, F. Davoli, S. Singh, The Potential Impact of Green Technologies in Next Generation Wireline Networks - Is There Room for Energy Saving Optimization?, IEEE Communications Magazine, vol. 49, no. 8, pp , Aug [5] R. Bolla, R. Bruschi, F. Davoli, F. Cucchietti, Energy Efficiency in the Future Internet: A Survey of Existing Approaches and Trends in Energy-Aware Fixed Network Infrastructures, IEEE Communications Surveys & Tutorials, vol. 13, no. 2, pp , 2nd Qr [6] L. Chiaraviglio, M. Mellia, F. Neri, Minimizing ISP Network Energy Cost: Formulation and Solutions, IEEE/ACM Transactions on Networking, vol. 20, no. 2, pp , April [7] A. Cianfrani, V. Eramo, M. Listanti, M. Polverini, An OSPF Enhancement for Energy Saving in IP Networks, Proc. IEEE INFOCOM 2011 Workshop on Green Communications and Networking, Shanghai China, April 2011, pp [8] J. C. Cardona Restrepo, C. G. Gruber, C. Mas Machuca, Energy Profile Aware Routing, Proc. GreenComm ICC Workshop, Dresden, Germany, June 2009.
13 [9] R. Bolla, R. Bruschi, F. Davoli, L. Di Gregorio, P. Donadio, L. Fialho, M. Collier, A. Lombardo, D. Reforgiato, T. Szemethy, "The Green Abstraction Layer: A Standard Power Management Interface for Next- Generation Network Device," IEEE Internet Computing, vol. 17, no. 2, pp March-April [10] R. Bolla, R. Bruschi, A. Carrega, F. Davoli, Green Networking with Packet Processing Engines: Modeling and Optimization", IEEE/ACM Transactions on Networking, 2013 (to appear); doi: /TNET Biographies Raffaele Bolla is a full professor of telecommunication networks at the University of Genoa, Italy, and principal investigator for the ECONET (low Energy COnsumption NETworks) project. raffaele.bolla@unige.it. Roberto Bruschi is a researcher with the National Inter-University Consortium for Telecommunications (CNIT), Genoa, Italy, the coordination assistant of the ECONET project, and the principal investigator of the GreenNet project. roberto.bruschi@cnit.it. Franco Davoli is a full professor of telecommunication networks at the University of Genoa, Italy, and is a member of the CNIT Scientific Board. franco.davoli@cnit.it. Pasquale Donadio is a system engineer at Alcatel-Lucent Italia, where he manages several research programs funded by the European Community. pasquale.donadio@alcatel-lucent.com Leonardo Fialho is a postdoctoral researcher in network energy efficiency at Dublin City University, Ireland. leonardofialho@gmail.com Martin Collier is a senior lecturer at Dublin City University, Ireland, where he runs the Switching and Systems Laboratory. martin.collier@dcu.ie. Alfio Lombardo is a full professor of telematics at the University of Catania, Italy, where he s responsible for the OpenLab. alfio.lombardo@dieei.unict.it. Diego Reforgiato Recupero is a postdoctoral researcher in network energy efficiency at the University of Catania, Italy. diegoref@dieei.unict.it. Vincenzo Riccobene is a Ph.D. student in network energy efficiency at the University of Catania, Italy. vincenzo.riccobene@dieei.unict.it. Tivadar Szemethy is chief architect for Netvisor in Hungary, which focuses on network and systems operation and management. tivadar.szemethy@netvisor.hu.
14 Network Control and Management Processes (e.g., Signalling Protocols) Layer 0: Logical Resources Layer 1: Device Root Layer 2: 1 st level HW Res. Layer 3: 2 nd level HW Res a Layer n: (n-1) th level HW Res m Convergence Layer 1...m
15 Layer 3: HW Components Layer 2: Layer 1: Device Layer 0: Logical Resources CLI/Routing Protocols eth1 eth2 eth40 DROP data-plane Entity 1 Entity 1.1 FE1 Entity 1.2 FE2 Entity 1.3 FE3
16 Consumption Consumption Consumption Standby Power Scaling Low Power Idle (LPI) AR z AR z-1 Adaptive Rate (AR) AR 2 AR 1 S k S 1 LPI v LPI 1 Wake-up time Wake-up time Max Rate
17 Consumption time AR 1 AR z LPI v AR z AR 2 LPI 1 AR AR 1 zarz-1 AR 2
18 Consumption Consumption Offered Load time AR 1 AR z LPI v AR z AR 2 LPI 1 AR AR 1 zarz-1 AR 2
19
20 Load Balancing IP routing Demultiplexing Interconnection (RX side) Interconnection (TX side)
21 Power Consumption [W] Power Consumption [W] Power Consumption [W] CLI eth1 eth2 eth40 Entity Autonomic 1 5 Autonomic 2 0 BAU Traffic Load [%] ) Control the number of FEs in standby & move traffic flows among FEs Autonomic Autonomic ) Decompose the energy Standby profiles from hardware 30 BAU entities to logical ones Traffic Load [%] 150 Autonomic 1 Autonomic 2 BAU Traffic Load [%] Entity 1.1 Entity 1.2 Entity 1.3 Interconnection (RX side) Interconnection (TX side)
22 S-PsS attributes. Parameter k Standby Power Saving Maximum Wake-Up time Maximum Sleeping time Wakeup Triggers Sleeping Timer P-PsS attributes. Parameter Description Identifier number of the current S-PsS. The power gain for being at S k with respect to the maximum power draw of the entity. The maximum time that is needed to wake-up from S k to S 0. The maximum time that is needed to go into S-PsS S k from active S 0. Set of flags that represents which triggers can be used to wake-up the entity. Bit 0: ; Bit 1: wake-onpacket; Bit 2: WoL. Other flags will be defined. Time of inactivity before the entity enters into S k in an autonomic way. Description General attr. AR Specific attr. LPI Specific attr. Autonomic attr. State Trans. attr. j Minimum Gain Maximum Throughput Maximum Throughput LPI Power Gain Power Packet LPI Wakeup Time LPI Sleeping Time LPI Power Autonomic Steps Autonomic Profile Curves Bit Transition PS PS Autonomic PS Service Interruption P-PsS Times Transition P-PsS Transition Service Interruption Times Identifier number of the current Power Scaling PsS. The minimum power gain in P j. with respect to the maximum power draw of the entity. This value is usually reached when the entity works at the maximum throughput in P j. The maximum throughput provided by the entity when working in P j. This value is expressed in packets per second. The maximum throughput provided by the entity when working in P j. This value is expressed in bits per second. The average power gain when the entity is idle in P j with respect to the maximum power draw of the entity. The maximum time that is needed to wake-up from the LPI mode upon packet arrival. The maximum time that is needed to enter into the LPI mode. The average power needed in P j when moving from idle to active modes and vice versa. Granularity used for describing the power and performance profile in autonomic power scaling primitives. This parameter represents the behavior of autonomic PS solutions. It is an array of depth STEPS (i.e., Throttling steps). Each array cell is composed by three values, indicating the offered load, the maximum consumption and the maximum packet service time at that load. Maximum service interruption time that may be incurred when the autonomic power scaling mechanisms switches the entity capacity. Array containing the maximum time intervals needed to move from P j to other P-PsS. The infinity value is used for indicating that a direct transition is not possible. Array containing the maximum service interruption times that may incur while moving from P j to other Power Scaling PsS. The infinity value is used for indicating that a direct transition is not possible.
Energy COnsumption NETworks (ECONET)
Energy COnsumption NETworks (ECONET) Prof. Raffaele Bolla CNIT University of Genoa Paris 25th Sept. 2012 The Project Motivations and Focus x 10 Greenhouse gas emission estimation according to GeSI Static
More informationlow Energy COnsumption NETworks
low Energy COnsumption NETworks (ECONET) Smart power management for fixed network devices ETSI Workshop on Energy Efficiency Genova 21st June 2012 The Project Motivations and Focus x 10 Static energy efficiency
More informationEvaluating the Energy-Awareness of Future Internet Devices
20 IEEE 2th International Conference on High Performance Switching and Routing Evaluating the Energy-Awareness of Future Internet Devices Raffaele Bolla DIST - University of Genoa Genoa, Italy raffaele.bolla@unige.it
More informationENERGY-EFFICIENT PROTOCOL IN OMNET++ SIMULATION ENVIRONMENT. Krzysztof Daniluk
159 ENERGY-EFFICIENT PROTOCOL IN OMNET++ SIMULATION ENVIRONMENT Krzysztof Daniluk Abstract: New idea for energy-efficient extension for OSPF protocol is presented. New concept is based on Dijkstra s algorithm,
More informationGreen Networking in Wired and Wireless Networks Bridging the Gap
low Energy COnsump1on NETworks Green Networking in Wired and Wireless Networks Bridging the Gap Franco Davoli Na9onal Inter University Consor9um for Telecommunica9ons (CNIT) and DIST University of Genoa
More informationThe Potential Impact of Green Technologies in Next-Generation Wireline Networks: Is There Room for Energy Saving Optimization?
ENERGY EFFICIENCY IN COMMUNICATIONS The Potential Impact of Green Technologies in Next-Generation Wireline Networks: Is There Room for Energy Saving Optimization? Raffaele Bolla and Franco Davoli, DIST-University
More informationGreen Network Technologies and the Art of Trading-off
IEEE INFOCOM 2011 Workshop on Green Communications and Networking Green Network Technologies and the Art of Trading-off Raffaele Bolla, Roberto Bruschi, Alessandro Carrega, and Franco Davoli Abstract In
More informationlow Energy COnsumption NETworks
E C O N E T low Energy COnsumption NETworks The aimed at studying and exploiting dynamic adaptive technologies (based on standby and performance scaling capabilities) for wired network devices that allow
More information5G and Beyond: Perspectives on Fixed/Mobile/Cloud/Fog Integration, Network Management and Control, and Services Deployment
5G and Beyond: Perspectives on Fixed/Mobile/Cloud/Fog Integration, Network Management and Control, and Services Deployment Raffaele Bolla *,, Roberto Bruschi, Franco Davoli *, * Telematics and Telecommunication
More informationGreen networking: lessons learned and challenges Prof. Raffaele Bolla CNIT/University of Genoa
Telecommunication s and Telematics Lab Green networking: lessons learned and challenges Prof. Raffaele Bolla raffaele.bolla@unige.it CNIT/University of Genoa Department of Naval, Electrical, Electronics
More informationDistributed Energy-Aware Routing Protocol
Distributed Energy-Aware Routing Protocol Erol Gelenbe and Toktam Mahmoodi Intelligent Systems & Networks Group Department of Electrical & Electronic Engineering Imperial College, London SW7 2BT, UK {e.gelenbe,t.mahmoodi}@imperial.ac.uk
More informationModeling Performance and Energy Efficiency of Virtualized Flexible Networks
In-Network Programmability for nextgeneration personal cloud service support A HOLISTIC, INNOVATIVE FRAMEWORK FOR THE DESIGN, DEVELOPMENT AND ORCHESTRATION OF 5G- READY APPLICATIONS AND NETWORK SERVICES
More informationNetwork softwarization, network management and energy efficiency: friends or foes?
Network softwarization, network management and energy efficiency: friends or foes? Raffaele Bolla, Roberto Bruschi, Franco Davoli TNT-Lab, DITEN - University of Genoa and CNIT University of Genoa Research
More informationSIMULATION OF ENERGY-AWARE BACKBONE NETWORKS
SIMULATION OF ENERGY-AWARE BACKBONE NETWORKS Ewa Niewiadomska-Szynkiewicz, Andrzej Sikora, Marcin Mincer, Piotr Arabas Institute of Control and Computation Engineering Warsaw University of Technology Nowowiejska
More information1 Energy Efficient Protocols in Self-Aware Networks
Energy Efficient Protocols in Self-Aware Networks Toktam Mahmoodi Centre for Telecommunications Research King s College London, London WC2R 2LS, UK Stanford NetSeminar 13 December 2011 1 Energy Efficient
More informationAn Open-Source Platform for Distributed Linux Software Routers
1 An Open-Source Platform for Distributed Linux Software Routers Raffaele Bolla DITEN, University of Genoa, Italy {name.surname}@unige.it Roberto Bruschi CNIT, Italy {name.surname}@cnit.it Abstract In
More informationIX: A Protected Dataplane Operating System for High Throughput and Low Latency
IX: A Protected Dataplane Operating System for High Throughput and Low Latency Belay, A. et al. Proc. of the 11th USENIX Symp. on OSDI, pp. 49-65, 2014. Reviewed by Chun-Yu and Xinghao Li Summary In this
More informationEfficient Power Management and Data Intensive Computer Systems in Computer Networks
Indian Journal of Science and Technology, Vol 8(12), 64835, June 2015 ISSN (Print) : 0974-6846 ISSN (Online) : 0974-5645 Efficient Power Management and Data Intensive Computer Systems in Computer Networks
More informationSoftware-Defined Networking (SDN) Overview
Reti di Telecomunicazione a.y. 2015-2016 Software-Defined Networking (SDN) Overview Ing. Luca Davoli Ph.D. Student Network Security (NetSec) Laboratory davoli@ce.unipr.it Luca Davoli davoli@ce.unipr.it
More informationEnergy Efficient Packet Processing Engine
Energy Efficient Packet Processing Engine 1 G.Joseph Antony Rosario, 2 Mr. S. Veerakumar 1 M.E Applied Electronics, Student, Department of ECE, Jaya Engineering College, Tamilnadu, India 2 Assistant Professor,
More informationExecuting Evaluations over Semantic Technologies using the SEALS Platform
Executing Evaluations over Semantic Technologies using the SEALS Platform Miguel Esteban-Gutiérrez, Raúl García-Castro, Asunción Gómez-Pérez Ontology Engineering Group, Departamento de Inteligencia Artificial.
More informationEnergy Efficient Mobile. Konstantinos Samdanis and Manuel Paul
Energy Efficient Mobile Backhaul: A BBF Initiative Konstantinos Samdanis and Manuel Paul June 2015 Outline Motivation Environmental and Cost Considerations Regulatory Policies Energy Efficient Mobile Backhaul
More informationCognitive augmented routing system and its standardisation path
Cognitive augmented routing system and its standardisation path ETSI Future Network Technologies Workshop Dimitri Papadimitriou, Bernard Sales Alcatel-Lucent, Bell Labs March, 2010 Self-Adaptive (top-down)
More informationHOW SDN AND NFV EXTEND THE ADOPTION AND CAPABILITIES OF SELF-ORGANIZING NETWORKS (SON)
WHITE PAPER HOW SDN AND NFV EXTEND THE ADOPTION AND CAPABILITIES OF SELF-ORGANIZING NETWORKS (SON) WHAT WILL YOU LEARN? n SON in the radio access network n Adoption of SON solutions typical use cases n
More informationIntroducing MESSIA: A Methodology of Developing Software Architectures Supporting Implementation Independence
Introducing MESSIA: A Methodology of Developing Software Architectures Supporting Implementation Independence Ratko Orlandic Department of Computer Science and Applied Math Illinois Institute of Technology
More informationSUMMERY, CONCLUSIONS AND FUTURE WORK
Chapter - 6 SUMMERY, CONCLUSIONS AND FUTURE WORK The entire Research Work on On-Demand Routing in Multi-Hop Wireless Mobile Ad hoc Networks has been presented in simplified and easy-to-read form in six
More informationA Path Decomposition Approach for Computing Blocking Probabilities in Wavelength-Routing Networks
IEEE/ACM TRANSACTIONS ON NETWORKING, VOL. 8, NO. 6, DECEMBER 2000 747 A Path Decomposition Approach for Computing Blocking Probabilities in Wavelength-Routing Networks Yuhong Zhu, George N. Rouskas, Member,
More informationEnergy efficiency management and environmental impact assessment for ICTs
http://eustandards.in/ Energy efficiency management and environmental impact assessment for ICTs Beniamino Gorini (ETSI TC-EE chairman) 2 Contents What is required to define sustainable ICT products/networks/services
More informationPerformance of Multihop Communications Using Logical Topologies on Optical Torus Networks
Performance of Multihop Communications Using Logical Topologies on Optical Torus Networks X. Yuan, R. Melhem and R. Gupta Department of Computer Science University of Pittsburgh Pittsburgh, PA 156 fxyuan,
More informationSeparation of concerns among application and network services orchestration in a 5G ecosystem
Separation of concerns among application and network services orchestration in a 5G ecosystem Panagiotis Gouvas, Anastasios Zafeiropoulos, Eleni Fotopoulou, Thanos Xirofotos Network Softwarization and
More informationTransport Area Working Group. Intended status: Standards Track. September 3, 2018
Transport Area Working Group Internet-Draft Intended status: Standards Track Expires: March 7, 2019 J. Saldana J. Fernandez Navajas J. Ruiz Mas University of Zaragoza September 3, 2018 Simplemux. A generic
More informationDomain Based Metering
Domain Based Metering Róbert Párhonyi 1 Bert-Jan van Beijnum 1 1 Faculty of Computer Science, University of Twente P.O. Box 217, 7500 AE Enschede, The Netherlands E-mail: {parhonyi, beijnum}@cs.utwente.nl
More informationBackground Brief. The need to foster the IXPs ecosystem in the Arab region
Background Brief The need to foster the IXPs ecosystem in the Arab region The Internet has become a shared global public medium that is driving social and economic development worldwide. Its distributed
More informationPERSONAL communications service (PCS) provides
646 IEEE/ACM TRANSACTIONS ON NETWORKING, VOL. 5, NO. 5, OCTOBER 1997 Dynamic Hierarchical Database Architecture for Location Management in PCS Networks Joseph S. M. Ho, Member, IEEE, and Ian F. Akyildiz,
More informationEuropean Network on New Sensing Technologies for Air Pollution Control and Environmental Sustainability - EuNetAir COST Action TD1105
European Network on New Sensing Technologies for Air Pollution Control and Environmental Sustainability - EuNetAir COST Action TD1105 A Holistic Approach in the Development and Deployment of WSN-based
More informationProgrammable BitPipe. Andreas Gladisch VP Convergent Networks and Infrastructure, Telekom Innovation Labs
Programmable BitPipe Andreas Gladisch VP Convergent Networks and Infrastructure, Telekom Innovation Labs 25.10.2012 How do you program a switch / router today? Vendor N SDK and API Vendor 3 Vendor 2 SDK
More informationThe UCD community has made this article openly available. Please share how this access benefits you. Your story matters!
Provided by the author(s) and University College Dublin Library in accordance with publisher policies., Please cite the published version when available. Title Implications of energy efficient Ethernet
More informationThe Energy Consumption of TCP
The Energy Consumption of TCP Raffaele Bolla DITEN University of Genoa Genoa - Italy raffaelle.bolla@unige.it Roberto Bruschi CNIT Genoa - Italy roberto.bruschi@cnit.it Paolo Lago DITEN University of Genoa
More informationService-Oriented Advanced Metering Infrastructure for Smart Grids
Journal of Energy and Power Engineering 5 (2011) 455-460 Service-Oriented Advanced Metering Infrastructure for Smart Grids S. Chen 1, J.J. Lukkien 1 and L. Zhang 2 1. Department of Mathematics and Computer
More informationEnd-to-end QoS negotiation in network federations
End-to-end QoS negotiation in network federations H. Pouyllau, R. Douville Avril, 2010 Outline Motivation for Network federations The problem of end-to-end SLA composition Scenario of composition and negotiation
More informationModelling direct application to network bandwidth provisioning for high demanding research applications
Modelling direct application to network bandwidth provisioning for high demanding research applications H. Wessing, Y. Yan and M. Berger Research Center COM Technical University of Denmark Bldg. 345V,
More information8. CONCLUSION AND FUTURE WORK. To address the formulated research issues, this thesis has achieved each of the objectives delineated in Chapter 1.
134 8. CONCLUSION AND FUTURE WORK 8.1 CONCLUSION Virtualization and internet availability has increased virtualized server cluster or cloud computing environment deployments. With technological advances,
More information6WINDGate. White Paper. Packet Processing Software for Wireless Infrastructure
Packet Processing Software for Wireless Infrastructure Last Update: v1.0 - January 2011 Performance Challenges for Wireless Networks As advanced services proliferate and video consumes an ever-increasing
More informationOn characterizing BGP routing table growth
University of Massachusetts Amherst From the SelectedWorks of Lixin Gao 00 On characterizing BGP routing table growth T Bu LX Gao D Towsley Available at: https://works.bepress.com/lixin_gao/66/ On Characterizing
More informationITU-T Y Next generation network evolution phase 1 Overview
I n t e r n a t i o n a l T e l e c o m m u n i c a t i o n U n i o n ITU-T Y.2340 TELECOMMUNICATION STANDARDIZATION SECTOR OF ITU (09/2016) SERIES Y: GLOBAL INFORMATION INFRASTRUCTURE, INTERNET PROTOCOL
More informationENERGY-EFFICIENT TRAFFIC MERGING FOR DATACENTER NETWORKS
ENERGY-EFFICIENT TRAFFIC MERGING FOR DATACENTER NETWORKS 1 RUSHIKESH MADHUKAR BAGE, 2 GAURAV DIGAMBAR DOIPHODE 1, 2 Graduate Student, Department Of Computer Science, Mumbai, Maharashtra, India 1 Email
More informationUsing Flexibility as a Measure to Evaluate Softwarized Networks
Chair of Communication Networks Department of Electrical and Computer Engineering Technical University of Munich Using Flexibility as a Measure to Evaluate Softwarized Networks Wolfgang Kellerer Technical
More informationCaching video contents in IPTV systems with hierarchical architecture
Caching video contents in IPTV systems with hierarchical architecture Lydia Chen 1, Michela Meo 2 and Alessandra Scicchitano 1 1. IBM Zurich Research Lab email: {yic,als}@zurich.ibm.com 2. Politecnico
More informationMobile Cloud Multimedia Services Using Enhance Blind Online Scheduling Algorithm
Mobile Cloud Multimedia Services Using Enhance Blind Online Scheduling Algorithm Saiyad Sharik Kaji Prof.M.B.Chandak WCOEM, Nagpur RBCOE. Nagpur Department of Computer Science, Nagpur University, Nagpur-441111
More informationReservation Packet Medium Access Control for Wireless Sensor Networks
Reservation Packet Medium Access Control for Wireless Sensor Networks Hengguang Li and Paul D Mitchell Abstract - This paper introduces the Reservation Packet Medium Access Control (RP-MAC) protocol for
More informationAbstract of the Book
Book Keywords IEEE 802.16, IEEE 802.16m, mobile WiMAX, 4G, IMT-Advanced, 3GPP LTE, 3GPP LTE-Advanced, Broadband Wireless, Wireless Communications, Cellular Systems, Network Architecture Abstract of the
More informationSOFTWARE ARCHITECTURE & DESIGN INTRODUCTION
SOFTWARE ARCHITECTURE & DESIGN INTRODUCTION http://www.tutorialspoint.com/software_architecture_design/introduction.htm Copyright tutorialspoint.com The architecture of a system describes its major components,
More informationEnergy Efficient Computing Systems (EECS) Magnus Jahre Coordinator, EECS
Energy Efficient Computing Systems (EECS) Magnus Jahre Coordinator, EECS Who am I? Education Master of Technology, NTNU, 2007 PhD, NTNU, 2010. Title: «Managing Shared Resources in Chip Multiprocessor Memory
More informationEnergy-aware Traffic Allocation to Optical Lightpaths in Multilayer Core Networks
Telecommunication Networks and integrated Services (TNS) Laboratory Department of Digital Systems University of Piraeus Research Center (UPRC) University of Piraeus Energy-aware Traffic Allocation to Optical
More informationPolicy-Based Context-Management for Mobile Solutions
Policy-Based Context-Management for Mobile Solutions Caroline Funk 1,Björn Schiemann 2 1 Ludwig-Maximilians-Universität München Oettingenstraße 67, 80538 München caroline.funk@nm.ifi.lmu.de 2 Siemens AG,
More informationTOSSIM simulation of wireless sensor network serving as hardware platform for Hopfield neural net configured for max independent set
Available online at www.sciencedirect.com Procedia Computer Science 6 (2011) 408 412 Complex Adaptive Systems, Volume 1 Cihan H. Dagli, Editor in Chief Conference Organized by Missouri University of Science
More informationDelayed reservation decision in optical burst switching networks with optical buffers
Delayed reservation decision in optical burst switching networks with optical buffers G.M. Li *, Victor O.K. Li + *School of Information Engineering SHANDONG University at WEIHAI, China + Department of
More informationETSI-AFI work on SDN-oriented Enablers for Customizable Autonomic Management & Control in Evolving and Future Networks
ETSI-AFI work on SDN-oriented Enablers for Customizable Autonomic Management & Control in Evolving and Future Networks ETSI 2013. All rights reserved ETSI AFI Activities & Roadmap Ranganai Chaparadza,
More informationEnable Infrastructure Beyond Cloud
Enable Infrastructure Beyond Cloud Tim Ti Senior Vice President R&D July 24, 2013 The Ways of Communication Evolve Operator s challenges Challenge 1 Revenue Growth Slow Down Expense rate device platform
More information2 Background: Service Oriented Network Architectures
2 Background: Service Oriented Network Architectures Most of the issues in the Internet arise because of inflexibility and rigidness attributes of the network architecture, which is built upon a protocol
More informationScalable Coding of Image Collections with Embedded Descriptors
Scalable Coding of Image Collections with Embedded Descriptors N. Adami, A. Boschetti, R. Leonardi, P. Migliorati Department of Electronic for Automation, University of Brescia Via Branze, 38, Brescia,
More informationDecentralized and Embedded Management for Smart Buildings
PROCEEDINGS OF THE WORKSHOP ON APPLICATIONS OF SOFTWARE AGENTS ISBN 978-86-7031-188-6, pp. 3-7, 2011 Decentralized and Embedded Management for Smart Buildings Giancarlo Fortino and Antonio Guerrieri DEIS
More informationNEC Virtualized Evolved Packet Core vepc
TE-524262 NEC Virtualized Evolved Packet Core vepc Design Concepts and Benefits INDEX Leading the transformation into Mobile Packet Core Virtualization P.3 vepc System Architecture Overview P.4 Elastic
More informationAlternative Network Deployments: Taxonomy, characterization, technologies and architectures
Alternative Network Deployments: Taxonomy, characterization, technologies and architectures draft-irtf-gaia-alternative-network-deployments-01 Authors: J. Saldana, University of Zaragoza, Ed. A. Arcia-Moret,
More informationSimple Quality-of-Service Path First Protocol and Modeling Analysis*
Simple Quality-of-Service Path First Protocol and Modeling Analysis* Lin Shen, Mingwei Xu, Ke Xu, Yong Cui, Youjian Zhao Department of Computer Science, Tsinghua University, Beijing, P.R.China, 100084
More informationInternet Research Task Force (IRTF) Category: Informational May 2011 ISSN:
Internet Research Task Force (IRTF) T. Li, Ed. Request for Comments: 6227 Cisco Systems, Inc. Category: Informational May 2011 ISSN: 2070-1721 Abstract Design Goals for Scalable Internet Routing It is
More informationPower Aware Hierarchical Epidemics in P2P Systems Emrah Çem, Tuğba Koç, Öznur Özkasap Koç University, İstanbul
Power Aware Hierarchical Epidemics in P2P Systems Emrah Çem, Tuğba Koç, Öznur Özkasap Koç University, İstanbul COST Action IC0804 Workshop in Budapest - Working Group 3 May 19th 2011 supported by TUBITAK
More informationA Methodology in Mobile Networks for Global Roaming
ORIENTAL JOURNAL OF COMPUTER SCIENCE & TECHNOLOGY An International Open Free Access, Peer Reviewed Research Journal Published By: Oriental Scientific Publishing Co., India. www.computerscijournal.org ISSN:
More informationBackground Brief. The need to foster the IXPs ecosystem in the Arab region
Background Brief The need to foster the IXPs ecosystem in the Arab region The Internet has become a shared global public medium that is driving social and economic development worldwide. Its distributed
More informationLoad Balancing with Minimal Flow Remapping for Network Processors
Load Balancing with Minimal Flow Remapping for Network Processors Imad Khazali and Anjali Agarwal Electrical and Computer Engineering Department Concordia University Montreal, Quebec, Canada Email: {ikhazali,
More informationArchitecture of Distributed Systems Component-based Systems
Architecture of Distributed Systems 2017-2018 Component-based Systems Original : J.J Lukkien Revision: R.H. Mak 25-Oct-17 Rudolf Mak TU/e Computer Science 2II45-CBSE Goals of this lecture Students have
More informationŁabiak G., Miczulski P. (IIE, UZ, Zielona Góra, Poland)
UML STATECHARTS AND PETRI NETS MODEL COMPARIS FOR SYSTEM LEVEL MODELLING Łabiak G., Miczulski P. (IIE, UZ, Zielona Góra, Poland) The system level modelling can be carried out with using some miscellaneous
More informationEnergy Efficient EE-DSR Protocol for MANET
Energy Efficient EE- Protocol for MANET 1 Mr. Prakash Patel, 2 Ms. Tarulata Chauhan 1 Department of Computer engineering, 1 LJ Institute of Technology, Ahmedabad, India 1 prakashpmp1990@gmail.com, 2 taruchauhan114@gmail.com
More informationImplementation of a Wake-up Radio Cross-Layer Protocol in OMNeT++ / MiXiM
Implementation of a Wake-up Radio Cross-Layer Protocol in OMNeT++ / MiXiM Jean Lebreton and Nour Murad University of La Reunion, LE2P 40 Avenue de Soweto, 97410 Saint-Pierre Email: jean.lebreton@univ-reunion.fr
More informationVERTICAL QOS MAPPING OVER WIRELESS INTERFACES
ACCEPTED FROM O PEN C ALL VERTICAL QOS MAPPING OVER WIRELESS INTERFACES MARIO MARCHESE AND MAURIZIO MONGELLI, UNIVERSITY OF GENOA Modern telecom networks are composed of different portions and technologies.
More informationUsing Low-power Modes for Energy Conservation in Ethernet LANs
Using Low-power Modes for Energy Conservation in Ethernet LANs Maruti Gupta Department of Computer Science Portland State University Portland, OR 9727 Email: mgupta@cs.pdx.edu Suresh Singh Department of
More informationTag Switching. Background. Tag-Switching Architecture. Forwarding Component CHAPTER
CHAPTER 23 Tag Switching Background Rapid changes in the type (and quantity) of traffic handled by the Internet and the explosion in the number of Internet users is putting an unprecedented strain on the
More informationSustNMS: Towards Service Oriented Policy-Based Network Management for Energy-Efficiency
SustNMS: Towards Service Oriented Policy-Based Network Management for Energy-Efficiency Carlos H. A. Costa, Marcelo C. Amaral, Guilherme C. Januário, Tereza C. M. B. Carvalho Escola Politécnica, University
More informationCommunication System Design Projects
Communication System Design Projects KUNGLIGA TEKNISKA HÖGSKOLAN PROFESSOR: DEJAN KOSTIC TEACHING ASSISTANT: GEORGIOS KATSIKAS Traditional Vs. Modern Network Management What is Network Management (NM)?
More informationProtocol Structure Overview of QoS Mapping over Satellite Networks
Protocol Structure Overview of QoS Mapping over Satellite Networks Mario Marchese, Maurizio Mongelli DIST - Department of Communication, Computer and System Sciences University of Genoa, Via Opera Pia
More informationSIMULATION ISSUES OF OPTICAL PACKET SWITCHING RING NETWORKS
SIMULATION ISSUES OF OPTICAL PACKET SWITCHING RING NETWORKS Marko Lackovic and Cristian Bungarzeanu EPFL-STI-ITOP-TCOM CH-1015 Lausanne, Switzerland {marko.lackovic;cristian.bungarzeanu}@epfl.ch KEYWORDS
More informationThe Role of SNMP in A General TMN Framework For Convergent Networks
The Role of SNMP in A General TMN Framework For Convergent Networks Wolfgang Haidegger Telecommunications Research Center (FTW) Vienna Donau-City-Str. 1, TechGate A-1220 Wien Austria Phone: +43 1 5052830-63
More informationMonitoring services on Enterprise Service Bus
Monitoring services on Enterprise Service Bus Ilona Bluemke, Marcin Warda Institute of Computer Science, Warsaw University of Technology, Nowowiejska 15/19, 00-665 Warsaw, Poland {I.Bluemke}@ii.pw.edu.pl
More informationPrecomputation Schemes for QoS Routing
578 IEEE/ACM TRANSACTIONS ON NETWORKING, VOL. 11, NO. 4, AUGUST 2003 Precomputation Schemes for QoS Routing Ariel Orda, Senior Member, IEEE, and Alexander Sprintson, Student Member, IEEE Abstract Precomputation-based
More informationNGSON: Features, State of the Art, and Realization
NEXT GENERATION SERVICE OVERLAY NETWORKS (NGSON) NGSON: Features, State of the Art, and Realization Seung-Ik Lee and Shin-Gak Kang, Standards Research Center, Electronics and Telecommunications Research
More informationA Novel Optimization Method of Optical Network Planning. Wu CHEN 1, a
A Novel Optimization Method of Optical Network Planning Wu CHEN 1, a 1 The engineering & technical college of chengdu university of technology, leshan, 614000,china; a wchen_leshan@126.com Keywords:wavelength
More informationThomas Lin, Naif Tarafdar, Byungchul Park, Paul Chow, and Alberto Leon-Garcia
Thomas Lin, Naif Tarafdar, Byungchul Park, Paul Chow, and Alberto Leon-Garcia The Edward S. Rogers Sr. Department of Electrical and Computer Engineering University of Toronto, ON, Canada Motivation: IoT
More informationInternet Traffic Characteristics. How to take care of the Bursty IP traffic in Optical Networks
Internet Traffic Characteristics Bursty Internet Traffic Statistical aggregation of the bursty data leads to the efficiency of the Internet. Large Variation in Source Bandwidth 10BaseT (10Mb/s), 100BaseT(100Mb/s),
More informationIPv6-based Beyond-3G Networking
IPv6-based Beyond-3G Networking Motorola Labs Abstract This paper highlights the technical issues in IPv6-based Beyond-3G networking as a means to enable a seamless mobile Internet beyond simply wireless
More informationA MAC Layer Abstraction for Heterogeneous Carrier Grade Mesh Networks
ICT-MobileSummit 2009 Conference Proceedings Paul Cunningham and Miriam Cunningham (Eds) IIMC International Information Management Corporation, 2009 ISBN: 978-1-905824-12-0 A MAC Layer Abstraction for
More informationKeeping the Connectivity and Saving the Energy in the Internet
IEEE INFOCOM 11 Workshop on Green Communications and Networking Keeping the Connectivity and Saving the Energy in the Internet Francesca Cuomo, Anna Abbagnale, Antonio Cianfrani, Marco Polverini DIET,
More informationTowards Integration of Slice Networking in NFV
Towards Integration of Slice Networking in NFV draft-gdmb-netslices-intro-and-ps-02 draft-galis-anima-autonomic-slice-networking-02 V3.0 30 th March 2017 Prof. Alex Galis a.gais@ucl.ac.uk;; http://www.ee.ucl.ac.uk/~agalis/
More informationOn the Impact of Route Processing and MRAI Timers on BGP Convergence Times
On the Impact of Route Processing and MRAI Timers on BGP Convergence Times Shivani Deshpande and Biplab Sikdar Department of ECSE, Rensselaer Polytechnic Institute, Troy, NY 12180 Abstract Fast convergence
More informationDesigning a System Engineering Environment in a structured way
Designing a System Engineering Environment in a structured way Anna Todino Ivo Viglietti Bruno Tranchero Leonardo-Finmeccanica Aircraft Division Torino, Italy Copyright held by the authors. Rubén de Juan
More informationDesign and Development of Carrier Assignment and Packet Scheduling in LTE-A and Wi-Fi
PhD Defense Design and Development of Carrier Assignment and Packet Scheduling in LTE-A and Wi-Fi Husnu S. Narman Outline Introduction Objective Multi-band in LTE-A Multi-band in Wi-Fi Conclusion Husnu
More informationApplication Oriented Networks: An SOA Perspective
Oriented s: An SOA Perspective www.thbs.com Introduction Service Oriented Architecture is the hot topic of discussion in IT circles today. So much so, in fact, that SOA is being seen by many as the future
More informationExtensions to RTP to support Mobile Networking: Brown, Singh 2 within the cell. In our proposed architecture [3], we add a third level to this hierarc
Extensions to RTP to support Mobile Networking Kevin Brown Suresh Singh Department of Computer Science Department of Computer Science University of South Carolina Department of South Carolina Columbia,
More informationCarrier SDN for Multilayer Control
Carrier SDN for Multilayer Control Savings and Services Víctor López Technology Specialist, I+D Chris Liou Vice President, Network Strategy Dirk van den Borne Solution Architect, Packet-Optical Integration
More informationITU-T Y Framework of multi-homing in IPv6-based NGN
INTERNATIONAL TELECOMMUNICATION UNION ITU-T Y.2052 TELECOMMUNICATION STANDARDIZATION SECTOR OF ITU (02/2008) SERIES Y: GLOBAL INFORMATION INFRASTRUCTURE, INTERNET PROTOCOL ASPECTS AND NEXT-GENERATION NETWORKS
More informationLow pass filter/over drop avoidance (LPF/ODA): an algorithm to improve the response time of RED gateways
INTERNATIONAL JOURNAL OF COMMUNICATION SYSTEMS Int. J. Commun. Syst. 2002; 15:899 906 (DOI: 10.1002/dac.571) Low pass filter/over drop avoidance (LPF/ODA): an algorithm to improve the response time of
More information