Transport Layer QoS Management for Wireless Multimedia Services

Transcription

1 Transport Layer QoS Management for Wireless Multimedia Services Abstract Vijay K. Madisetti and Antonios D. Argyriou Soft.Networks, LLC and Georgia Tech 09/13/ :28:13 AM EST We introduce a new transport-layer technology, VoMo, optimized for delivery of multimedia services over future wireless/wired networks that is based on innovations at the transport layer of the network. The Quality of Service (QoS) implications of this technology appear promising, and its close relationship to emerging IETF protocols, such as the Stream Control Transmission Protocol (SCTP) are expected to ensure that it can be quickly introduced into practice, in applications such as CDNs, VoIP, and Video over IP/Cable/Wireless networks. Introduction Future wireless all-ip networks, while offering the promise of exciting broadband applications, are expected to consist of several, potentially incompatible, wireless access technologies that would be offered by a number of competing service providers. Once past the access stage, the Internet is still expected to be the main traffic backbone. The diversity of access technologies, however, may affect the Quality of Service (QoS). These QoS issues become even more important when the user may possibly want to switch from one access technology to the other that may either is priced differently or provides better service or it is available in a place where another one is not. Additional issues arise when considering the widely differing types of services that the user may use: streaming media, real-time communications, interactive communications, VoIP, Multimedia-commerce are representative of some of them. Each of these services imposes its distinct QoS requirements. These QoS demands from multimedia traffic are compounded in the case of a wireless network, where new problems arise due to the implied mobility of the users as well as due to the nature of the current IP protocols that support IP-based mobility, combined with a lossy and interrupt/outage-prone nature of the communications channel. QoS requirements can be achieved by various techniques that span multiple layers of the network OSI hierarchy. However, we believe that transport layer is an oft overlooked area with respect to the promise that it offers for improvement of mobility management, QoS, security, and throughput. Our intention is to optimize all this functionality at the transport layer in order to create a framework for easier and robust application development. Additionally a transport layer-based approach has to be able to interoperate with future QoSaware backbone networks which will be primarily based on MPLS and DiffServ technologies. 1. Multimedia transmission over wireless channels Multimedia transmission over wireless links presents a great challenge for the internet engineering community today. Since the wireless multimedia services experience is judged according to the delivered quality, we are focusing primarily in the analysis of the QoS requirements of multimedia applications. We also identify the problems that occur in the case where these kinds of applications are deployed in wireless systems and how QoS is affected.

2 A. Types of wireless multimedia traffic Multimedia traffic may be generally categorized into streaming, interactive (real-time), or simply non-real-time. With streaming multimedia, constant transfer rates are of primary importance. This includes applications that are related to audio and video streams. So, brief disruptions in the transfer rate become noticeable by the user, resulting in jittering pictures or stuttering during sound playback. However, some initial delay is acceptable. Another category is the interactive multimedia. In this case we are concerned with real-time data transfers and temporal fluctuations. So, while one may tolerate minor errors during content transfer, long delays and jitter are usually unacceptable from QoS considerations. Voice over IP and video telephony are representative of the applications that belong to this category. New exciting applications such as M-commerce would also fit this category. Non-real-time multimedia traffic such as Video on Demand (VoD), are concerned primarily about throughput and not with delay or jitter. B. Wireless traffic problems As noted in the preceding section, each type of multimedia service has different QoS requirements and attempts to solve a differing set of problems. These problems, specially applied to wireless networks, may be summarized as Limited Bandwidth: Wireless bandwidth is known to be limited and the wireless medium itself is subject to problems such as multi-path fading and noise susceptibility. 2. High packet loss rate and bit error rate. 3. Need for Mobility management: The need for uninterrupted communication in the presence of roaming users implies handoff procedures between various wireless access points. However, this procedure comes with the price; lost packets and interruption of an ongoing season. 4. Heterogeneous 3G infrastructure (GPRS, UMTS, WLAN, Bluetooth). Each technology cannot be deployed everywhere and so wireless content delivery systems may have to be transparent to various underlying technologies. 5. Inadequate performance of Transport Layer protocols (e.g., TCP/UDP). The old transport layer in IP networks based on TCP and UDP, that was designed for data networks, is not expected to perform satisfactorily in the wireless content delivery network scenario. 6. QoS guarantees. Due to the aforementioned problems, providing Quality of Service guarantees is not expected to be easy, and new and adaptable QoS service models are expected to be needed. 7. Wireless Security. Advanced authentication and encryption techniques that combine performance with low power requirements are needed to ensure widespread acceptance of wireless multimedia services. 2. QoS in Mobile and Wireless Content Delivery Networks We will now present a quick overview of these approaches that intend to improve QoS at the network, transport and application layers, primarily related to QoS in the access networks. The list is not exhaustive, and pointers to further ongoing work can be found in the references section. Table 1 depicts a summary of work at the application, transport, network and

3 data/physical link layers. We also note that we are not only focusing on approaches that intend to improve QoS directly, but also on approaches that achieve it indirectly (e.g., through efficient mobility management mechanisms). Approaches such as DiffServ and MPLS are not mentioned further here, since these technologies target QoS at the Internet backbone. Data Link / Physical HiperLAN Network Mobile IP Micromobilty: HAWAII, CIP Neighbor Casting [20] Transport AOTP [1] Application SIP [2] Rate adaptation Approaches Table 1. Various QoS/Mobility Approaches A. Application Layer QoS Mobility with SIP Currently, promising approaches that provide mobility management for media data over wireless links are based on SIP (Session Initiation Protocol) [1] and Mobile IP standards [2][3][4]. SIP is not a QoS management protocol. However there are ongoing efforts [6] that are positioning SIP to be used as a protocol that is responsible for mobility management at the application layer, with an additional objective to improve QoS for real-time & non-real-time multimedia applications. SIP, according to [6], provides session, terminal, personal and service mobility. SIP relies on the usage of SIP's registration mechanism for providing terminal and personal mobility. Every user has a URI, which could be an address (e.g., burdell@gatech.edu). When the user moves to a new place, it registers its new position with the SIP registrar server, so that the SIP registrar server knows the user's current position. This kind of registration can provide user-level mobility. In order to achieve terminal mobility a SIP mobility implementation could poll the operating system (OS) in order to find out if handoff took place so that the SIP will register with the new SIP registrar after handoff [6]. SIP based mobility, thus, offers attractive benefits when used in mobile multimedia applications. However, there are some inherent problems with this approach that make the adoption of this scheme difficult. For example it cannot handle mid-call subnet changes, since it is an application layer solution. This is where it requires the support of a lower level mobility

4 protocol, e.g., Mobile IP. One other important issue is that of inter-operability with Mobile IP. Home Agent and Foreign Agent registrations in mobile IP serve the same purpose with the SIP REGISTER messages and their joint deployment becomes problematic. B. Network Layer QoS The Network Layer is the primary OSI layer for enforcing QoS policies in computer networks. This is because the core network components (routers), that play the major role in network performance, operate at the network layer (IP). A number of network layer approaches exist that intend to improve mobility management at the IP layer, and thus the QoS indirectly. Mobile IP is the most well-known mobility management solution. However, handoff delay, and overheads of Mobile IP's triangular routing, triangular registration, and IP encapsulation are major issues that present a bottleneck for mobile IP to become a wide spread acceptable solution for real-time interactive multimedia communications over the wired or wireless IP network. Various mobile IP [2] modifications like HAWAII [5], Cellular IP [4], Domain based approaches [3] are focusing on improving mobility management especially in the case of frequent handoffs. However they do not explicitly address QoS-related problems. More information concerning these approaches can be found in the references section. An interesting approach found in [3] attempts to improve the QoS by providing a more scalable reservation system for wireless applications through localized RSVP messages. In this approach RSVP is used in such a way that when a mobile node moves to a new point of attachment new PATH and RESV messages are send only locally, between the mobile node and an anchor point, creating thus a hierarchy in the QoS reservation system. This approach has the potential of improving significantly resource reservation time. C. Transport Layer QoS Transport layer is a neglected area concerning QoS related research. However, there are a few notable exceptions. In [7] a new experimental transport protocol, the Application-Oriented Transport Protocol (AOTP) is presented. AOTP is above IP and provides transport services with new functionality added specifically, to trade off reliability, throughput and/or jitter, in order to support applications of the upper layer with the required QoS. AOTP is claimed to provide adjustable partially reliable service, priority-based error recovery strategies and dynamic playback management. This approach however represents a "multimedia-aware" protocol that is targeting wire-line networks without any kind of provisioning for wireless systems. 3. Transport Layer QoS Solutions As mentioned earlier, older transport layer protocols, such as TCP and UDP, do not appear to be able to fully meet the stringent QoS requirements for interactive multimedia services, and we now examine a new transport layer protocol, called Stream Control Transport Protocol (SCTP). SCTP is a new IETF transport level protocol [8] that offers a number of advanced transport layer services. Primary purpose for the design of this protocol was to be able to transfer reliably SS7 signaling messages over IP-based networks. However it soon proved to be not

5 only an application specific protocol but it could also overthrow TCP since it has a number of advanced novel features that we will now describe. SCTP introduces the idea of multi-homing, where a host has multiple interfaces and IP addresses by which it is reachable. An association between two endpoints can exist between any of these addresses. If one of the paths that correspond to one address fails then an alternative can be used without interrupting the connection between the endpoints. The two endpoints can monitor the status of the paths by sending a special kind of SCTP message called the Heartbeat. Primary goal of the above protocol property was error resilience. Additionally SCTP provides the ability to maintain multiple streams of messages inside a single association. This makes possible to maintain a sequence of messages only per stream basis (partial in-sequence delivery) [9] thus reducing unnecessary head-of-line blocking between streams of messages that are independent. Another important feature is the distinction between the delivery mechanism and reliable datagram transfer. This provides a more flexible usage of the protocol so that is adapted to the specific needs of the application using it. It is, for example, possible for a scenario where one application requires partial ordering of the delivered datagrams, while another could be satisfied with reliable transfer that does not imply any kind of sequencing. Table 4 summarizes the primary differences between SCTP and TCP. SCTP Stream Based Multi-homing Association has multiple streams 4-Way connection establishment Reliability & delivery mechanism are separate TCP Byte Oriented One connection Single data flow 3-Way handshake Unified mechanism Table 2. SCTP / TCP Comparison An SCTP implementation, fully compatible with the RFC, appears to be able to perform better than TCP even in the case of a wireless system. Even though SCTP does not incorporate any novel features particularly suited for wireless systems, it is a powerful new transport technology that has a number of advantages over TCP in this regard. Multi-homing can greatly improve performance: even a standard SCTP stack operating in a mobile computer with more than one network interface cards can significantly increase data transfer reliability [10]. Moreover SCTP supports IPv6 and it can it can operate at the same time by using IPv4-IPv6

6 addresses. This feature is of importance since IPv6 is expected to soon replace the older IP version. Secondary SCTP Path Multihomed Host Multihomed Host Primary SCTP Path Figure 3. SCTP multihoming 4. Efficient Multimedia Services over Networks VoMo Enhancements to SCTP: the VoMo platform The VoMo or Voice over Mobile IP platform enhances the base SCTP protocol at the transport layer in many ways, and relies on the following technologies: An Intelligent Address Distribution (IAD) mechanism A Neighborhood Based Mobility (NBM) mechanism A QoS-aware Transport Layer with following capabilities: o Path Quality Monitoring, o Network Status Estimation o Efficient stream & physical path management o Rate Control & Content Adaptation mechanisms Load-Balancing & Bandwidth Aggregation mechanisms These technologies ensure that VoMo can support highly optimized and efficient implementation of wireless multimedia services.

7 SCTP Stream Based Only HEARTBEAT [8] Unified congestion control QoS aware VoMo Path monitoring (delay, throughput, e.t.c) Bandwidth Aggregation Congestion control per link Mobility management capable Enhanced Socket API: Various data classes Enhanced services: Content adaptation Table 4. SCTP / VoMo Comparison Operation of VoMo The VoMo platform consists of two distinct phases that are related to mobility management. The initialization process phase is performed at each subnet even when no mobile nodes are inside a cell. The connection process phase handles all the necessary steps that a mobile node (MN) has to go through in order to connect to the wireless network. Phase 1: Initialization of the network All contiguous cells to the currently active cell, will be automatically included in the neighborhood_list, ensuring that any handoff will be smoother based on the movement of the mobile node (MN). Phase 2: The connection process Step 1: Mobile Node enters a subnet and sends a registration request. Step 2: A central server (e.g., DHCP) provides a list_of_addresses based on the neighborhood_list. Step 3: The MN uses as its primary Care-of-Address one of the list_of_addresses that corresponds to its current point of attachment When the MN moves to another point of attachment (or a Base Station, BS) two scenarios arise: Step 4a: When the MN movement is inside the subnet, it MUST switch to another IP listed in its list_of_addresses according to movement detection information. Step 4b: When MN is moving to a new subnet, it sends a registration request to the other subnet requesting new list_of_addressees

8 Step 5: The MN decides when it has changed its BS via information on movement detection obtained from the L2 or L3 protocol layers. Step 6: Upon completion of its decision, the MN, having moved to a new subnet discards the old list_of_addresses and starts using a new_primary_address from a newly allocated list_of_addresses. Correspondent Host Home Network We assume that the Correspondent Node has already discovered the Mobile Node s current address Step 1:Correspondent Host invites the Mobile Node to a conference with SIP or H.323 and SCTP Foreign Network Step 2: Mobile Host accepts the invitation and can send immediately media data with U-SCTP The SCTP connection is preserved during the whole time the two nodes communicate Mobile host The Mobile Host sends media data from the path that satisfies the QoS requirements Figure 5. VoMo connection establishment QoS Features Supporting Multimedia Services In the VoMo platform, multimedia data sets are distinguished at the transport layer according to the classes these data sets are assigned to by the application layer, creating an opportunity to provide Quality of Service (QoS). The classes of data sets are assumed to match those proposed by the Universal Mobile Telephone System (UMTS) [11]: e.g., conversational class, streaming class, interactive class, and background class. Adopting the application classes as the above is not restrictive for any kind of application layer protocol and may offer the right abstraction for manipulating data at the modified SCTP layer. The QoS features intelligently route data to outgoing appropriate interfaces based on network status information, available network resources, application layer requirements, and application bit-rate (BER) demands. More details on operation of QoS features may be found in Reference [12]. Content adaptation for wireless multimedia traffic Due to wireless and mobility issues, network bandwidth and noise characteristics of paths between various communications endpoints are expected to vary over time. VoMo offers the

9 ability to adapt the quality of the content being transmitted based on the channel or communications capabilities available to the service. Load-Balancing The VoMo platform implements load balancing at the transport layer. While the foundations rely on the multi-homing technology of SCTP, several new features are provided. The user can use a new Stream Ordering (SO) service where he can specify an ordering in the stream data delivery. This can be useful in the case where the user wants to transfer a single file and wants to split it into streams for transmission across various available interfaces. VoMo handles this case by using the SO service to split data into streams, and to tailor the transmission by sending the streams out according to the available bandwidth at each interface. The VoMo load-balancing mechanism operates in close cooperation with the QoS features. The QoS module makes decisions that satisfy the application requirements and the Load- Balancing mechanism is invoked when the use of more than one interface is needed. Mobile IP Bandwidth Packet Loss Delay Improved Jitter No VoMo High Low Low Low SIP High High Table 6. Advantages of VoMo compared to other mobile wireless approaches (for QoS metrics) A. Expected impact & Experimental Results The VoMo system is expected to have the following impact: Reduced handoff delay Infrastructure for handoff is simpler Allows efficient QoS management and load balancing at the transport layer Implements content class-driven path selection Provides improved error resilience.

10 Performance degradation due to handoff effects will be reduced and lost packets due to cell handoff are expected to be eliminated. By establishing a multi-homed connection near the cell boundaries, the mobile device, when experiencing handoff, will gradually switch to another interface which corresponds to the new cell. For instance, let us suppose that packets are lost during handoff while they are being transmitted from the old cell. In this case, when the mobile device enters the new cell and realizes it did not receive an acknowledgement, it will retransmit the missing data. Figure 7. Total receiver throughput under base SCTP and VoMo For the sake of completeness we present simulation results that were performed by using the NS-2 network simulator [13]. Part of VoMo functionality has been implemented as a pluggable module to this simulator. Figure 5 shows simulation results depicting the total throughput at the receiver using base SCTP and VoMo approaches. In this specific setup we considered two hosts communicate through 2 links of 500Kbps each and delay of 100ms. Additionally we used reliable transfer for SCTP and VoMo Figure 6 presents the total receiver throughput at the receiver for the same topology except that we added packet loss rate equal to 2%. This is rather increased but depicts clearly performance under high loss links which is usually the case in wireless links. However, the total amount of used bandwidth is far more increased in the case of VoMo. Moreover in the case where the two links have far different loss conditions VoMo selects the a path with a very small packet loss rate so that it can provide an outgoing flow with more stable rate. This is particularly important for delay-sensitive applications.

11 Figure 8. Total receiver throughput under base SCTP and VoMo with increased loss conditions 5. Examples / Case Studies A. Application Example (MPEG-4) MPEG-4 [14] video coding and MPEG-7 [14] content retrieval technology are expected to be the driving forces that will deliver multimedia content over the next generation mobile devices. These applications are well suited for advanced transport services offered by the VoMo platform as will be described in the paragraphs that follow. The MPEG-4 standard is actually a set of tools that are available to the developer, and can be used according to required coding requirements, decoder complexity, and data format, to name a few options. MPEG-4 is characterized by profiles, which are sets of tools that provide a specific functionality. For example the Simple Profile is uses the H.263 video conferencing standard. This profile is tailored for low bit-rate applications which would usually include handheld wireless devices. Other complex MPEG-4 profiles exist which include synchronized interactive environments with a number of arbitrary-shaped video objects, and associated 2-D and 3-D vector graphics. A typical MPEG-4 streaming video platform performs rate control offering VBR and CBR services [14]. However, this approach is network agnostic and is based on efficient implementations of MPEG-4 coding standards and the underlying media protocols RTP/RTCP/RTSP. We now describe how MPEG-4 streaming can be implemented efficiently through VoMo.

12 Secondary SCTP Path Mobile Multihome dhost Host Primary SCTP Path Figure 9. Wireless Access Scenario Transmitting MPEG-4 encoded video with VoMo The MPEG-4 Simple Profile produces a base layer (BL) and several enhancement layers (EL) which can enhance the base layer description quality. Since the BL is the most important description it should receive the best protection, preferably using forward error correction (FEC). VoMo allows the handling of each layer according to its importance. The base layer can be assigned to a VoMo stream that corresponds to a highly reliable path. The MPEG-4 Core Visual Profile adds support for the encoding of objects that are arbitrarilyshaped. If this profile were used, MPEG-4 encoding is performed at an object by object basis. Important objects may be encoded preferentially, and VoMo would consider these MPEG-4 objects as individually different transport entities. This implies that the delivery of each one of them may be handled separately by a suitable mapping procedure between these objects and transport layer streams. Interactive object manipulation only affects those streams that are carrying these object, while the rest of the bit stream stays unaffected.

13 Enhancement Bit Stream Basic Layer Switch Enhancement Bit Stream Basic Layer Figure 10. In the first case, transmission of enhancement layers under the same link may lead to two streams competing for it, thus degrading basic stream QoS. VoMo uses non-overlapped paths thus not disturbing base layer. 6. Summary/Conclusions We propose a new technology for management of QoS for multimedia content delivery over wired and wireless networks that is based on innovations at the transport layer, utilizing the best of breed protocols, such as IETF s Stream Control Transmission Protocol (SCTP) and novel methods for handling mobile wireless users that appear superior to those utilized by Mobile IP. Statement on IP Rights: Some material described in this white paper may be part of pending patent applications through the Office of Technology Licensing at Georgia Institute of Technology, Atlanta, Georgia, 30332, USA. Additional Reading & References [1] M. Handley et al.,"sip: Session Initiation Protocol", RFC 2543, March [2] C. Perkins, "IP Mobility Support", RFC 2002, October [3] K. Kim et al. "Domain Based Approach for QoS Provisioning over Mobile IP", in Proc. Of ICON 2001, Bangkok, Thailand. [4] Campbell, A.T., et al "Design, implementation, and evaluation of cellular IP", IEEE Personal Communications, pp , August [5] Ramjee, R. et al, "HAWAII: A Domain-Based Approach for Supporting Mobility in Wide- Area Wireless Networks", IEEE/ACM Transactions on Networking, pp , June [6] Ashutosh Dutta, et al "Application Layer Mobility Management Scheme for Wireless Internet," 3Gwireless 2001, San Francisco, pp. 7, May [7] V. Tsaoussidis, S. Wei, "QoS Management at the Transport Layer", ITCC 2000, Las Vegas, Nevada. [8] R. R. Stewart, Q. Xie, K. Morneault, C. Sharp, H. J. Schwarzbauer, T. Taylor, I. Rytina, M. Kalla, L. Zhang, and V. Paxson, "Stream Control Transmission Protocol", RFC 2960, October 2000.

14 [9] R. Stewart, M. Ramalho et al., "SCTP Partial Reliability Extension", <draft-stewart-tsvwgprsctp-00.txt >, May [10] M. Riegel et al. "Mobile SCTP", <draft-riegel-tuexen-mobile-sctp-00.txt>, February [12] V. Madisetti, A. Argyriou, "An Transport Layer QoS Algorithm", <draft-madisetti-argyriouqos-sctp-00.txt >, July 25, [11] [13] [14]