CHAPTER1 INTRODUCTION 1.1 OVERVIEW Many applications would require fast data transfer in high-speed wireless networks nowadays. Howe ver, due to it s conservative congestion control algorithm, Transmission Control Protocol (TCP) cannot effe ctively utilize the network capac ity in lo ssy wireless networks. Here a re ceiver-assisted congestion 1
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
of these acknowledges are taken as a packet loss and reduces its congestion
window. This technique is very efficient in a traditional wired network.
Unfortunately, packet loss in a wireless network may also be due to
transmission problems such as a high link error probability, fading, and
interference. Therefore, packet loss is no longer an appropriate indication for
network congestion. With the wrong constructed information, TCP may
reduce its congestion window unnecessarily, resulting in poor
performance of wireless networks.
Another problem of TCP is its poor capability to utilize the network
bandwidth efficient ly, especially in networks with a high bandwidth delay product
(BDP). Bandwidth-delay product refers to the p r o du c t o f a data link's capacity(in b it s p e r s ec o n d ) and its e n d - to - e n d d e la y ( in seconds). The result, an amount
of data measured in bits (or b y t e s ) , is equivalent to the maximum amount of data
on the network circuit at any given time, i.e. data that has been transmitted but
not yet received. Sometimes it is calculated as the data link's capacity times its
r o u n d t r ip t im e . T he standard TCP congestion avoidance algorithm employs
an additive increase and multiplicative decrease (AIMD) scheme.
When there is no packet loss detected, the congestion window (cwnd)
is increased by one maximum segment size (MSS) every round-trip time (RTT).
Otherwise the TCP sender reduces cwnd by half if the packet loss is detected by
three duplicate ACKs or reduces cwnd to one if the packet loss is detected by
timeout. In a high-speed network with a large RTT, TCP requires a very large
window to efficiently utilize the net work resource. What is worse, upon a
retransmission timeout, the sender has to wait for a time duration before sending
any new packets, and this waiting t ime greatly reduces the TCP throughput.
Unlike regular TCP where the receiver only performs flow control, here
it is allowed the receiver to participate in congestion control. A timer is used at the
receiver to time the arrival of the next packet and to therefore detect a packet
P2P networking is the utilization of the relatively powerful computers (PCs) for more
than just client-based computing tasks. The modern PC has a very fast processor, vast
memory, and a large hard disk, none of which are being fully utilized when
performing common computing tasks. The modern PC can easily act as both a client
and server for many types of applications.
2.2 BINARY INCREASE CONGESTION CONTROL (BIC) FOR FASTLONG-DISTANCE NETWORKS
High-speed networks with large delays present a unique environment where TCP
may have a problem utilizing the full bandwidth. Several congestion control
proposals have been suggested to remedy this problem. The existing protocols
consider mainly two properties: TCP friendliness and bandwidth scalability. That is, a protocol should not take away too much bandwidth from standard TCP flows while
utilizing the full bandwidth of high-speed networks. This paper presents another
important constraint, namely, RTT (round trip time) unfairness where competing
flows with different RTTs may consume vastly unfair bandwidth shares. Existing
schemes have a severe RTT unfairness problem because the congestion window
increase rate gets larger as the window grows – ironically the very reason that makes
them more scalable. RTT unfairness for high-speed networks occurs distinctly with
drop tail routers for flows with large congestion windows where packet loss can be
highly synchronized. After identifying the RTT unfairness problem of existing
protocols, this paper presents a new congestion control scheme that alleviates RTT
unfairness while supporting TCP friendliness and bandwidth scalability.
2.3 FAST TCP: MOTIVATION, ARCHITECTURE, ALGORITHMS,PERFORMANCE
Control algorithm for high-speed long-latency networks, from design to
implementation. The approach taken by FAST TCP to address the four difficulties, at
both packet and flow levels, which the current TCP implementation has at large
windows. The architecture and summarize some of the algorithms implemented in
our prototype. The equilibrium and stability properties of FAST TCP.
Upon a timeout event (whether it is detected by the sender s timer or
informed by the receiver s ACK), the sender will decrease the congestion window
to one in consideration that the network is in congestion, and it will have some
time to recover. If congestion is mitigated after one RTT, the sender will
recover/adjust the congestion window in the next window by using the receiver
advertised window.
During fast retransmission, the sender sets the congestion window size
to the lesser value of the receiver advertised window size and the current size.
Since the packet loss may also indicate congestion, we should reduce the
congestion window. On the other hand, if the congestion window is less than
the receiver advertised window, the network may only be in a mild congestion.Therefore, it is unnecessary for the sender to reduce the congestion window.
In a normal state, when the sender receives an ACK, it will compare
the current congestion window with the receiver advertised congestion window.
If the receiver advertised window is larger than the sender s congestion
window, and the difference is larger than a predefined threshold, the sender
will set the congestion window size to the receiver advertised window.
Otherwise, the sender will ignore the receiver advertised congestion window by
just performing the additive increase mechanism.
3.2 ALGORITHM
i) When an ACK is sent, if rcv.rtt is zero, let rcv.rtt equal 1 and record the
corresponding sequence (rtseq) of data packet (equals to the sum of the
ACK sequence and the current congestion window.). Otherwise, just send
the ACK.
ii) If a data packet with a larger sequence number than rtseq is arrived in
order and the new measured packet inter-arrival interval is larger than two
times of pre-estimated packet inter-arrival interval, set the new measured
avoidance. The detail of the subsequent behavior of TCP depends on the version
of TCP.
4.1.1 SLOW-START FOR CONGESTION CONTROL
Slow-start is part of the congestion control strategy used by
TCP, the data transmission protocol used by many Internet applications, such as
HTTP and Secure Shell. Slow- start is used in conjunction with other algorithms
to avoid sending more data than the network is capable of transmitting, that is,
network congestion. Slow-start is one of the algorithms that TCP uses to control
congestion inside the network. It is also known as the exponential growth phase.
During the exponential growth phase, Slow-start works byincreasing the TCP congestion window each time the acknowledgment is
received. It increases the window size by number of segments acknowledged.
This happens until either an acknowledgment is not received for some segment or
a predetermined threshold value is reached. If a loss event occurs, TCP assumes
this it is due to network congestion and takes steps to reduce the offered load on
the network. Once a loss event has occurred or the threshold has been reached,TCP enters the linear growth (congestion avoidance) phase. At this point, the
window is increased by 1 segment for each RTT. This happens until a loss event
occurs.
4.1.2 BASIC SLOW-START FOR CONGESTION CONTROL ALGORITHM
The algorithm begins in the exponential growth phase initially with a
congestion window size (cwnd) of 1 or 2 segments and increases it by 1 Segment
Size (SS) for each ACK received. This behavior effectively doubles the window
size each round trip of the network. This behavior continues until the congestion
window size (cwnd) reaches the size of the receiver's advertised window or until
a loss occurs.
When a loss occurs half of the current cwnd is saved as a Slow Start13
When loss is detected, the policy is changed to be one of multiplicative decreasewhich is to cut the congestion window in half after loss. The result is a saw tooth
behavior that represents the probe for bandwidth. A loss event is generallydescribed to be either a timeout or the event of receiving 3 duplicate ACKs. Alsorelated to TCP congestion control is the slow start mechanism. Other policies or algorithms for fairness in congestion control are Additive Increase AdditiveDecrease (AIAD), Multiplicative Increase Additive Decrease (MIAD) and
Here in this work, in a wireless network, a sender and receiver
combined congestion control mechanism. The receiver estimates a congestion
window deemed to be appropriate from the measured bandwidth and RTT, andthen advertises the window size (feeds this information back) to the sender. The
sender then adjusts its congestion window according to the advertised window
of the receiver. Through this receiver-assisted method, the sender can increase
the congestion window quickly to the available bandwidth, thus improving the
network utilization. On the other hand, when timeout happens, the receiver can
feed this information quickly back to the sender to relieve the impact of
timeout to TCP performance. Compared to other receiver-centric mechanisms
or mobile-host- centric mechanisms, this mechanism has a simple
implementation with little modification required for the sender and receiver.
Moreover, when only the receiver end protocol is modified without changing the
sender functions, it can still improve the network throughput and fairness
performance.
This system can again be improved in the case of a lossy wireless
network such as a wireless LAN (WLAN) system, IEEE 802.11, in mult i-hop
scenarios using the medium access control (MAC) layer protocol in the place of
the TCP congestion control. It can be shown how its aggressive behavior can
throttle the spatial reuse and reduce bandwidth efficiency. An adaptive,
layer-2, distributed coordination scheme for 802.11 using explicit MAC
feedback is then proposed to pace the transmissions on adjacent nodes, thereby
assisting the MAC protocol to operate around its saturation state while
minimizing resource contention. Again, the sender can increase the congestion
window quickly to the available bandwidth, thus improving the network