K. Wang, S.K. Tripathi, Mobile-end transport protocols: an alternative to TCP/IP overwerSk() links, in: Proceedings of the IEEE Infocom#98, San Francisco, CA, USA, 1998, pp. 1046--1053.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....is often used for end to end error control at the transport layer. TCP, however, performs poorly over wireless links since it considers packet losses due to the radio channel as congestion losses and invokes congestion control mechanisms that significantly reduce the throughput [1] 2] 3] 4] [5], 6] To avoid such performance degradation, link layer solutions can be used to obtain local data reliability and make the wireless link appear to TCP as a more reliable link, although with a longer and variable delay. We consider an ARQ (Automatic Repeat reQuest) scheme at the link layer, and ....

K.-Y. Wang and S.K Tripathi, "Mobile-end transport protocol: an alternative to TCP/IP over wireless links," in Proc. of INFOCOM,

....sender and the base station and the other between the base station and the mobile receiver. The losses that are not related to congestion are recovered by the connection between the base station and the mobile host, and hence hidden from the fixed sender. I TCP [2] MTCP [3] M TCP [4] and METP [5] are examples of this approach. Some of these approaches do not maintain the end to end semantics of TCP. These protocols may require state to be maintained and packets to be buffered at the base station. In the TCP aware link layer protocols, the link layer is aware of the semantics of the TCP ....

K.-Y. Wang and S. K. Tripathi, "Mobile-end transport protocol: An alternative to TCP/IP over wireless links," IEEE INFOCOM'98, vol. 3, p. 1046, 1998.

....have suggested TCP enhancements and other mechanisms to accommodate wireless domain characteristics. They can be classified into several categories: Split the end to end TCP connection into wired and wireless parts to implement separate transmission protocols optimized for each domain [3] [16] [34] Enable network routers to generate explicit congestion or packet corruption messages to inform the sender to take relevant recovery actions [35] 36] Rely on link layer packet recovery mechanisms on the wireless link that automatically retransmits lost packets to preserve the TCP ....

K. Y. Wang and S. K. Tripathi, "Mobile-End Transport Protocol: An Alternative to TCP/IP Over Wireless Links", IEEE INFOCOM Proceedings, Vol. 3, 1998, pp. 1046-1053.

....the connection between the fixed host and the mobile host to be split into two separate connections at the base station one connection between the sender and the base station and the other between the base station and the receiver. The solutions such as I TCP [5] MTCP [4] M TCP [9] METP [14] come under this category. End To End protocols: In an end to end approach the sender and receiver adapt a mechanism to handle all possible packet losses. The solutions such as TCP New Reno [2] Fast Retransmission [6] Selective Acknowledgement [3] 11 ] SMART based Acknowledgement [ 11 ] ....

....link layer solutions will assure the high bit error rates seen in mobile environment. This protocol performs well in situations where the sender has a continuous stream of data to send. In these circumstances disconnections due to hand offs are handled efficiently. 4.1.4. Mobile end TCP (METP) [14] The Mobile end protocol hides the wireless link losses from the sender by replacing the TCP over the wireless link by a simple protocol with smaller headers if the link is the last hop along a data path. This protocol exploits the link layer acknowledgements and retransmissions to quickly ....

K.Wang and S. K. Tripathi, "Mobile-End Transport Protocol: An alternative to TCP/2 over wireless links", in IEEE Infocom, March 1998. Discusses a split-connection approach for TCP over wireless links.

....its mechanisms to familiarize the readers of the issues being dealt with by the proposed extensions to TCP. Readers familiar with concepts of TCP may skip this TCP for Wireless Networks 2 section without any loss of continuity. In Section 3, we have presented an overview of the splitconnection [6, 7, 8, 9, 11] approaches to solving the problem and have discussed their merits and drawbacks. Sections 4 and 5 talk about the TCP aware [12, 13, 14, 15] and TCP unaware [16] approaches respectively. Section 6 summarizes various Explicit Notification mechanisms [18, 5, 19] In Sections 7 and 8, we discuss ....

....Also, losses over wireless links tend to be bursty and frequent because of their vulnerability to interference and host mobility. The TCP sender assumes these data losses are due to congestion in the network and backs off resulting in performance degradation. To overcome these problems, Wang et al. [9] have proposed a split connection approach called Mobile End Transport Protocol (METP) The protocol replaces TCP IP over wireless link with a simpler protocol that uses smaller headers provided that the wireless link happens to be the last hop in the connection. Thus, the functionalities needed ....

[Article contains additional citation context not shown here]

K. Wang, and S. K. Tripathi, "Mobile-end transport protocol: An alternative to TCP/IP over wireless links," IEEE Infocom, pp. 1046-1053, March 1998.

....second TCP connection from base station to mobile host is optimized, whereas in the following scheme METP, the TCP is completely removed in the second connection. The connection between base station and mobile host is treated as interaction between transport layer and application layer. METP In [8], the authors have proposed a Mobile End Transport Protocol (METP) an alternative approach TCP over wireless networks. The basic idea is to remove TCP and replace it with a simpler protocol over wireless link; thereby we can eliminate the performance degradation of TCP due to wireless link ....

....hosts. The TCP IP stack of the hosts in the fixed network should not require modifications to improve the performance. In freezeTCP [4] and delayed dupacks [11] no modifications are required in the TCP IP stack in the hosts on the fixed network. On the other hand schemes such as SNOOP [8], I TCP [6] MTCP [7] MTCP [9] ESBN [2] require modifications in the base station are required. Maintaining the TCP end to end semantics. In most of the schemes the TCP semantics is violated, I TCP [6] and MTCP [7] violates the TCP s true end to end signaling by involving the intermediate ....

[Article contains additional citation context not shown here]

Kuang-Yeh Wang, Tripathi, S.K, "Mobile-end transport protocol: an alternative to TCP/IP over wireless links", INFOCOM '98, Seventeenth Annual Joint Conference of the IEEE Computer and Communications Societies. Proceedings. IEEE, Volume: 3, 1998 Page(s): 1046-1053 vol.3.

....la conveniencia de utilizar modelos indirectos de comunicacin aislando los problemas en una de las partes de la comunicacin, sustituyendo TCP en la parte problemtica por otro protocolo diseado ad hoc. Las simulaciones que se llevarn a cabo como parte del proyecto servirn de aval a estas ideas. En [Wang et al. 1998] se recoge un estudio que se asemeja bastante al del proyecto. Estudia entornos distintos de acceso a datos a travs de redes inalmbricas (concretamente acceso desde una red local inalmbrica WaveLAN a otra a travs de una red WAN almbrica) pero de igual manera basa las soluciones en el uso de ....

....este elemento. En [Rivanedeyra et al. 1998a] se propone que el intermediario sea un elemento al que llaman GSN (Gateway Support Node) cumpliendo tareas adicionales de ISP. De esta manera la divisin de mundos es entre el mundo GSM e Internet, no entre el mundo inalmbrico y el almbrico como propone [Wang et al. 1998], al colocar el intermediario en las estaciones base (Base Stations) de la red celular. La arquitectura Mowgli, presentada en [Kojo et al. 1995] propone situar el intermediario en el servidor de acceso a Internet (ISP) haciendo de esta manera que la divisin sea entre la red de acceso y la red ....

[Article contains additional citation context not shown here]

Kuang-Yeh Wang, Satish K.Tripathi (1998). Mobile-End Transport Protocol: An Alternative to TCP/IP Over Wireless Links. IEEE INFOCOM98, San Francisco, California, March 1998. URL: http://www.cs.umd.edu/users/kwang

....I TCP does not guarantee smooth hando since packets maybe lost while states are being transferred from the previous BS. A reliable transfer of states from BSs that are robust to varying migration scenario is presented in [95] The split connection approach was further improved by Wang et al. [96] and Brown et al. 97] Both Wang et al. and Brown et al. terminates TCP connections at the BS or supervisor host (SH) Then, a protocol (e.g. M TCP [97] which is optimized for the wireless link is used on the last hop. Caceres et al. 87] suggested fast retransmission after hando . After a MH ....

....it may cause high number of packets to be bu ered in the surrounding BSs. Therefore, it is bene cial if SmartBu er is incorporated into the snoop protocol to overcome errors at the wireless link as well as retransmission timeout due to high hando latency. The work presented by Wang et al. [96] is similar to SmartBu er. The following highlight the main di erences: SmartBu er does not require the splitting of TCP connections. This results in simpler solution because states (e.g. states allocated due to socket calls) pertaining to the connections do need to be transfered to the new BS ....

[Article contains additional citation context not shown here]

K.-Y. Wang and S. K. Tripathi, \Mobile-end transport protocol: An alternative to tcp/ip over wireless links," in IEEE INFOCOM'98, vol. 3, p. 1046, 98.

....College Station, TX 77843 3112, E mail: vaidya cs.tamu.edu, Phone: 409 845 0512, Fax: 409 847 8578 Microsoft Corporation, Redmond, WA. Mehta performed this work at the Texas A M University. Sun Microsystems Laboratories, Palo Alto Sun Microsystems Laboratories, Palo Alto [17] 19] 21] [25], 26] 27] However, many of these schemes require TCP specific actions on the part of intermediate nodes on the path from TCP sender to TCP receiver. Such schemes are sometimes referred as being TCP aware [3] The objective of this paper is to investigate an alternative TCP unaware technique ....

....are often mobile. To our knowledge, Caceres et al. 6] were the first to consider the impact of mobility on TCP performance, and suggest a mechanism to improve the throughput. Several other researchers have also considered techniques designed to take mobility into account [2] 6] 5] 17] [25]. IV. THE SNOOP SCHEME Because the proposed TCP unaware scheme attempts to imitate the TCP aware Snoop scheme [4] we first describe Snoop briefly. For the discussion below, consider the system shown in Figure 1. Consider TCP data transfer from node S to the wireless host WH, through the base ....

K. Wang and S. K. Tripathi, "Mobile-end transport protocol: An alternative to TCP/IP over wireless links," in IEEE Infocom, pp. 1046--1053, March 1998.

....error rate, when compared to wired links. Because of its inability to distinguish between packet losses due to congestion and due to transmission errors, TCP performs poorly on wireless links [2, 3] Several interesting approaches have already been proposed to improve TCP performance over wireless [2, 4, 3, 5, 8, 10, 15, 17, 19, 21, 25, 26, 27]. However, many of these schemes require TCP specific actions on the part of intermediate nodes on the path from TCP sender to TCP receiver. Such schemes are sometimes referred as being TCP aware [3] The objective of this paper is to investigate an alternative TCP unaware technique that can ....

....since wireless hosts are often mobile. To our knowledge, Caceres et al. 6] were the first to consider the impact of mobility on TCP performance, and suggest a mechanism to improve the throughput. Several other researchers have also considered techniques designed to take mobility into account [2, 6, 5, 17, 25]. 4 The Snoop Scheme Because the proposed TCP unaware scheme attempts to imitate the TCP aware Snoop scheme [4] we first describe Snoop briefly. For the discussion below, consider the system shown in Figure 1. Consider TCP data transfer from node S to the wireless host WH, through the base ....

K. Wang and S. K. Tripathi, "Mobile-end transport protocol: An alternative to TCP/IP over wireless links," in IEEE Infocom, pp. 1046--1053, March 1998.

....the BS can send Zero Window Adjustment (ZWA) to freeze the sender and preserve the congestion window as well as prevent a potential timeout. The METP is a split connection protocol that replaces the TCP IP protocol stack over the wireless interface with a simpler protocol with smaller headers [5]. In order to set up a communication with a fixed host, the routing and flow control functionality of the mobile host is undertaken by the BS (this is further facilitated by the one hop distance between the BS and the mobile node) An interesting feature of METP is that it exploits a link layer ....

K. Wang, S. Tripathi, Mobile-end transport protocol: an alternative to TCP/IP over wireless links, Proceedings IEEE Infocom (1998).

....during connection lifetime. In case the mapping changes, for example due to the host moving and attaching to another point in the network, the connection is hard to maintain. A large body of research exists that addresses this problem by adapting existing transport protocols to host mobility [2, 5, 24, 19]. We advocate a transport protocol which preserves and exploits the independence of the endpoint name space, as provided by the CBN, from the host address name space. We call the class of such protocols autonomous transport protocols (ATP) In an autonomous protocol, mobility can be trivially ....

....performance improvements. For example, a split for localized recovery might be useful where endto end mechanisms for error recovery would be too expensive, e.g. by avoiding a source based retransmission that consumes resources at intermediate nodes (power, bandwidth) Several mobile TCP schemes [2, 24], have also suggested a similar approach, trading endto end semantics for better performance over a hybrid wired wireless network. However, they are limited to one permanent, static split, at the boundary of the wired network. The remainder of the paper is structured as follows. We start with ....

[Article contains additional citation context not shown here]

K. Wang and S. K. Tripathi, \Mobile-End Transport Protocol: An Alternative to TCP/IP Over Wireless Links," INFOCOM (3), 1998.

No context found.

K. Wang, S.K. Tripathi, Mobile-end transport protocols: an alternative to TCP/IP overwerSk() links, in: Proceedings of the IEEE Infocom#98, San Francisco, CA, USA, 1998, pp. 1046--1053.

No context found.

K. Wang and S. K. Tripathi, "Mobile-end transport protocol: An alternative to TCP/IP over wireless links," in IEEE INFOCOM, vol. 3, Mar. 1998, pp. 1046--1053.

No context found.

K.-Y. Wang and S. K. Tripathi, "Mobile-end transport protocol: An alternative to TCP/IP over wireless links," IEEE INFOCOM'98, vol. 3, p. 1046, 1998.

No context found.

K. Y. Wang and S. K. Tripathi, "Mobile-End transport Protocol: An alternative to TCP/IP Over Wireless Links," Proceedings of INFOCOM, vol. 3, pp. 1046--1053, 1998.

No context found.

K. Y. Wang and S. K. Tripathi, "Mobile-End transport Protocol: An alternative to TCP/IP Over Wireless Links," Proceedings of INFOCOM, vol. 3, pp. 1046--1053, 1998.

No context found.

Kuang-Yeh Wang, Satish K.Tripathi (1998). MobileEnd Transport Protocol: An Alternative to TCP/IP Over Wireless Links. IEEE INFOCOM98, San Francisco, California, March 1998. URL: http://www.cs.umd.edu/users/kwang

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC