1 / 41

UNIT III

UNIT III . MOBILE TRANSPORT LAYER. INTRODUCTION. Supporting mobility only on lower layers up to the network layer is not enough to provide mobility support for applications.

rahn
Download Presentation

UNIT III

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. UNIT III MOBILE TRANSPORT LAYER

  2. INTRODUCTION • Supporting mobility only on lower layers up to the network layer is not enough to provide mobility support for applications. • Most applications rely on a transport layer, such as TCP (transmission control protocol) or UDP (user datagram protocol) in the case of the internet. • Two functions of the transport layer in the internet are checksumming over user data  multiplexing/demultiplexing of data from/to applications.

  3. The main difference between UDP and TCP is that TCP offers connections between two applications. • Within a connection TCP can give certain guarantees, such as in-order delivery or reliable data transmission using retransmission techniques. • TCP has built-in mechanisms to behave in a ‘network friendly’manner. • For example, TCP encounters packet loss, it assumes network internal congestion and slows down the transmission rate

  4. UDP do not behave in network friendly manner it does not aalyze the congestion and continue to sent the packet into an already congested network.

  5. Transport Layer

  6. Traditional TCP

  7. Characteristics of TCP • TCP is reliable. • TCP incorporates congestion control mechanism • TCP incorporates end-to-end flow control mechanism.

  8. Congestion control functionality of TCP is provided by five main algorithm • Congestion control • Slow start • Congestion avoidance • Fast retransmit • Fast recovery

  9. Congestion control • TCP is designated for fixed networks with fixed end systems. • Reason for packet loss in fixed network is a temporary overhead in transfer path. • In case of congestion network take decision to drop some packets. • Pack lost in not intimated by receiver instead it send acknowledgement upto missing packet . • Sender notice the ack is missing for lost packet and it assume packet is lost due to congestion. • To reduce congestion TCP slows down the transmission

  10. Slow Start or Exponential Start

  11. This method double the congestion window every time the ack come back and take on round trip time (RTT). • This is called Exponential growth of congestion window in slow start mechanism. Example • For new connection CWND is initialized to MSS (Max Segment Size) & Slow start Threshold (SST). • Sender generated exponential growth by increasing the CWND value by 1 MSS every time a transmitted. segment acknowledged • CWND =CWND+MSS

  12. Fast Retransmit/Recovery missing packet

  13. Fast retransmit/fast recovery • Two things lead to a reduction of the congestion threshold. • One is a sender receiving continuous acknowledgements for the same packet. • This informs the sender of about following things • One is that the receiver got all packets up to the acknowledged packet in sequence. • In TCP, a receiver sends acknowledgements only if it receives any packets from the sender. • Receiving acknowledgements from a receiver also shows that the receiver continuously receives something from the sender. • The gap in the packet stream is not due to severe congestion, but a simple packet loss due to a transmission error. • The sender can now retransmit the missing packet(s) before the timer expires. • This behavior is called fastretransmit

  14. The receipt of acknowledgements shows that there is no congestion to justify a slow start. • The sender can continue with the current congestion window. • The sender performs a fast recovery from the packet loss. • This mechanism can improve the efficiency of TCP dramatically. • The other reason for activating slow start is a time-out due to a missing acknowledgement.

  15. Congestion Avoidance • Congestion avoidance algm a retransmission timer expiring or reception of duplicate ACK can signal the sender that a network congestion situation is going on. • Sender set the transmission windowone half of the current window size ,but to atleast two segments. • If congestion timeoutcongestion window is reset to one segemnt which automatically put sender into Slow start mode. • If Congestion is due indicated by Fast retransmit & Fast recovery Algm is started

  16. Implications on Mobility • Slow start one of the most useful mechanism in fixed networks it decreases the efficiency of TCP used together with mobile receivers or senders. • Since slow start is used wrong assumptions during congestion. • Wireless link experience less error rates compares to fixed fiber or copper links.

  17. Influences of Wireless/mobility on TCP-mechanisms

  18. Classical TCP Solutions • Indirect TCP • Mobile TCP • Snooping TCP • Fast retransmit/Fast Recovery • Transmission/Time-out Freezing • Selective Retransmission • Transaction-Oriented TCP

  19. Indirect TCP developed to overcome one of the following • Poor Performance of TCP in wireless links • Inability of TCP to get fixed with a wireless part I-TCP segments TCP connection into two: • Fixed Part • Wireless Part

  20. Classical TCP ImprovementsEarly approach: Indirect TCP I

  21. I-TCP socket and state migration mobile hostaccess point2

  22. Indirect TCP II

  23. Snooping TCP • One of the drawbacks of I-TCP is the segmentation of the single TCP connection into two TCP connections. • This loses the original end-to-end TCP semantic. • The following TCP enhancement works completely transparently and leaves the TCP end-to-end connection intact. • The main function of the enhancement is to buffer data close to the mobile host to perform fast local retransmission in case of packet loss. • A good place for the enhancement of TCP could be the foreign agent in the Mobile IP

  24. Early approach: Snooping TCP I

  25. Snooping TCP II

  26. Advantages: • The end-to end TCP semantic is preserved • Correspondent host does not need to be changed: most of the enhancements are in the foreign agent • It does not need a hnadover of state as soon as a the mobile host moves to another foreign sgent. Disadvantages: • Snooping TCP does not isolate the behaviour of wireless link as well as I-TCP. • Using negative ACK between the FA and MH assumes additional mechanisms on MH.

  27. Mobile TCP Two important reasons for packet drop in wirelesss link and mobility are, • Due to handover or higher bit error rates. • Due to occurrence of Lengthy or frequent disconnection. Main objective of M-TCP is to improve overall throughtput to lower the delay to maintain end to end semantics and to provide a more efficient handover.

  28. Early approach: Mobile TCP

  29. Persistent Mode • Nature of sender to be same state even after disconnection from receiver for any time. • In this mode sender will not try to send any data

  30. Advantages: • M-TCP maintain the TCP end to end Semantics • SH does not send any Ack itself ,but forward the Acks from MH. Disadvantages: • M-TCP assumes low bit error rates which lead packet loss on wireless link due to bit errors. • A modified TCP on wireless link not only requires modifications to the MH protocol software but also new network elements like Bandwidth manager.

  31. Fast retransmit/fast recovery

  32. Transmission/time-out freezing

  33. Selective Retransmission TCP

  34. Transaction oriented TCP

  35. Comparison of different approaches for a “mobile” TCP

  36. TCP over 2.5/3G wireless networks • The current internet draft for TCP over 2.5G/3G wireless networks describes a profile for optimizing TCP over today’s and tomorrow’s wireless WANs such as GSM/GPRS, UMTS, or cdma2000. • The following characteristics have to be considered when deploying applications over 2.5G/3G wireless links: Data rates: • While typical data rates of today’s 2.5G systems are 10–20 kbit/s uplink and 20–50 kbit/s downlink. • 3G and future 2.5G systems will initially offer data rates around 64 kbit/s uplink and 115–384 kbit/s downlink. • Typically, data rates are asymmetric as it is expected that users will download more data compared to uploading. • To support multiple users Mobile communications within a radio cell, a scheduler may have to repeatedly allocate and deallocate resources for each user. • This may lead to a periodic allocation and release of a high-speed channel.

  37. Latency: • All wireless systems comprise elaborated algorithms for error correction and protection, such as forward error correction (FEC), check summing, and interleaving. • The current GPRS standard specifies an average delay of less than two seconds for the transport class with the highest quality Jitter: • Wireless systems suffer from large delay variations or ‘delay spikes’. • Reasons for sudden increase in the latency are: link outages due to temporal loss of radio coverage, blocking due to high-priority traffic, or handovers. Packet loss: • Packets might be lost during handovers or due to corruption. • However, recovery at the link layer appears as jitter to the higher layers

  38. Based on these characteristics, suggests the following configurationparameters to adapt TCP to wireless environments: Large windows: • TCP should support large enough window sizes based on the bandwidth delay product experienced in wireless systems. • With the help of the windows scale option (RFC 1323) and larger buffer sizes this can be accomplished Limited transmit: • This mechanism, is an extension of Fast Retransmission/Fast Recovery and is particularly useful when small amounts of data are to be transmitted (standard for, e.g., web service requests).

  39. Large MTU: • The larger the MTU (Maximum Transfer Unit) the faster TCP increases the congestion window. • Link layers fragment PDUs for transmission anyway according to their needs and large MTUs may be used to increase performance. • IPv6 should be used to employ larger segment sizes instead of assuming the small default MTU. Selective Acknowledgement (SACK): • SACK allows the selective retransmission of packets and is almost always beneficial compared to the standard cumulative scheme.

  40. Explicit Congestion Notification (ECN): • ECN allows a receiver to inform a sender of congestion in the network by setting the ECN-Echo flag on receiving an IP packet that has experienced congestion. • This mechanism makes it easier to distinguish packet loss due to transmission errors from packet loss due to congestion. Timestamp: • TCP connections with large windows may benefit from more frequent RTT samples provided with timestamps by adapting quicker to changing network conditions. • With the help of timestamps higher delay spikes can be tolerated by TCP without experiencing a spurious timeout. • The effect of bandwidth oscillation is also reduced.

  41. No header compression: • As the TCP header compression mechanism does not perform well in the presence of packet losses this mechanism should not be used.

More Related