Presentation is loading. Please wait.

Presentation is loading. Please wait.

TELE202 Lecture 8 Congestion control 1 Lecturer Dr Z. Huang Overview ¥Last Lecture »X.25 »Source: chapter 10 ¥This Lecture »Congestion control »Source:

Similar presentations


Presentation on theme: "TELE202 Lecture 8 Congestion control 1 Lecturer Dr Z. Huang Overview ¥Last Lecture »X.25 »Source: chapter 10 ¥This Lecture »Congestion control »Source:"— Presentation transcript:

1 TELE202 Lecture 8 Congestion control 1 Lecturer Dr Z. Huang Overview ¥Last Lecture »X.25 »Source: chapter 10 ¥This Lecture »Congestion control »Source: chapter 12 ¥Next Lecture »Internet Protocols (1) »Source: chapter 15

2 TELE202 Lecture 8 Congestion control 2 Lecturer Dr Z. Huang What is congestion? ¥Congestion »Congestion occurs when the number of packets being transmitted through the network approaches the packet handling capacity of the network »Effect: Too many packets are present in the subnet, a situation which causes performance degradation »Similar to road congestion

3 TELE202 Lecture 8 Congestion control 3 Lecturer Dr Z. Huang Queues in network ¥Data network is a network of queues »Packets arriving are stored at input buffers »Routing decision made »Packet moves to output buffer »Packets queued for output transmitted as fast as possible –Statistical time division multiplexing »If packets arrive to fast to be routed, or to be output, buffers will fill »Finite queues mean data may be lost, can discard packets »Generally 80% utilization is critical

4 TELE202 Lecture 8 Congestion control 4 Lecturer Dr Z. Huang Effects of Congestion ¥Ideal performance »Ideal assumes infinite buffers and no overhead

5 TELE202 Lecture 8 Congestion control 5 Lecturer Dr Z. Huang Effects of Congestion ¥Practical performance »Buffers are finite »Overheads occur in exchanging congestion control messages

6 TELE202 Lecture 8 Congestion control 6 Lecturer Dr Z. Huang Congestion control ¥Congestion control aims to keep number of packets below level at which performance falls off dramatically ¥Mechanisms »Backpressure »Choke packet »Traffic shaping/traffic policing »Implicit congestion signaling »Explicit congestion signaling »Load shedding

7 TELE202 Lecture 8 Congestion control 7 Lecturer Dr Z. Huang Backpressure ¥If a node becomes congested it can slow down or halt flow of packets from other nodes ¥May mean that other nodes have to apply control on incoming packet rates ¥Propagates back to source ¥Can restrict to logical connections generating most traffic ¥Used in connection oriented that allow hop by hop congestion control (e.g. X.25) ¥Not used in ATM nor frame relay ¥Only recently developed for IP ¥Disadvantage »May be too slow to react to congestion situations, because it takes a long time to propagate congestion information back to the source

8 TELE202 Lecture 8 Congestion control 8 Lecturer Dr Z. Huang Choke Packet ¥Choke packets are used as messengers to pass congestion information ¥Idea »Each router monitors the utilization of its output line. »If the utilization is above a threshold, the output line enters a “warning” state »When in a warning state, the router sends a choke packet back to the source host, giving it the destination found in the packet »The original packet is tagged so that the routers along the path will not generate choke packets for the packet »When the source host gets the choke packet, it is required to reduce the traffic sent to the specified destination by X percent. ¥ICMP source quench in Internet »From router or destination »Source cuts back until no more source quench message »Sent for every discarded packet, or anticipated

9 TELE202 Lecture 8 Congestion control 9 Lecturer Dr Z. Huang Traffic shaping ¥Traffic shaping is about regulating the average rate of data transmission »There is a contract between a customer and a carrier »The carrier monitors traffic flow from the customer according to the contract, a procedure called traffic policing ¥Leaky bucket algorithm »Imagine a bucket with a small hole in the bottom »No matter at what rate water enters the bucket, the outflow is at a constant rate »Once the bucket is full, any additional water entering it spills over the sides and is lost »The same idea can be applied to the packets »Each host is connected to the network by an interface containing a leaky bucket

10 TELE202 Lecture 8 Congestion control 10 Lecturer Dr Z. Huang Traffic shaping ¥Problem with leaky bucket »enforces a rigid output pattern at the average rate, no matter how bursty the traffic is »For some applications, it is better to allow the output to speed up when large bursts arrive. ¥Token bucket algorithm »Every ∆T second a token is generated and put into the bucket »When a packet is to be transmitted, a token must be captured and destroyed »If there is no token in the bucket, the algorithm can not transmit any packet »Idle hosts can save up permission (tokens) to send bursts later. The maximum number of tokens allowed to be saved is n »Token bucket algorithm does not discard packets

11 TELE202 Lecture 8 Congestion control 11 Lecturer Dr Z. Huang Congestion Signaling ¥Implicit Congestion Signaling »Transmission delay may increase with congestion »Packet may be discarded »Source can detect these as implicit indications of congestion »Useful on connectionless (datagram) networks »e.g. IP based –TCP includes congestion and flow control based on delay - see chapter 17 »Used in frame relay LAPF ¥Explicit Congestion Signaling »Network alerts end systems of increasing congestion »End systems take steps to reduce offered load »Backwards –Congestion avoidance in opposite direction to packet is required »Forwards –Congestion avoidance in the same direction as packet is required »Used in Frame Relay and ATM

12 TELE202 Lecture 8 Congestion control 12 Lecturer Dr Z. Huang Categories of Explicit Signaling ¥Binary »A bit set in a packet indicates congestion »Used in Frame Relay ¥Credit based »Indicates how many packets source may send »Common for end to end flow control ¥Rate based »Supply explicit data rate limit »e.g. ATM

13 TELE202 Lecture 8 Congestion control 13 Lecturer Dr Z. Huang Load shedding ¥Last resort! ¥When routers are being inundated by packets that they can not handle, they just throw them away ¥Dropping packets randomly may not reduce congestion. »Dropping one packet may cause several packets coming back! »Remember the error control algorithm! ¥Select the right packets to drop is very important »For file transfer, old packet is worth more than a new one (wine) »For multimedia, a new packet is more important than an old one (milk) ¥Senders’ co-operation is needed »Senders must mark their packets in priority classes to indicate how important they are »A full frame is more important than a difference frame in compressed video transmission »The routers can first drop packets from the lowest class, then the next lowest class, and so on.

14 TELE202 Lecture 8 Congestion control 14 Lecturer Dr Z. Huang More issues in congestion control ¥Traffic Management »Fairness »Quality of service –May want different treatment for different connections »Reservations –e.g. ATM –Traffic contract between user and network ¥Congestion Control in Packet Switched Networks »Send control packet to some or all source nodes –Requires additional traffic during congestion »Rely on routing information –May react too quickly »End to end probe packets –Adds to overhead »Add congestion info to packets as they cross nodes –Either backwards or forwards

15 TELE202 Lecture 8 Congestion control 15 Lecturer Dr Z. Huang Summary ¥Congestion »Effect of congestion ¥Congestion control »Backpressure »Choke packet »Traffic shaping/traffic policing »Implicit congestion signaling »Explicit congestion signaling »Load shedding ¥Issues in congestion control


Download ppt "TELE202 Lecture 8 Congestion control 1 Lecturer Dr Z. Huang Overview ¥Last Lecture »X.25 »Source: chapter 10 ¥This Lecture »Congestion control »Source:"

Similar presentations


Ads by Google