Lecture, November 27, 2002

1
Lecture, November 27, 2002

• TCP
• Other Internet Protocols Internet Traffic
• Scalability of Virtual Circuit Networks
• QoS

2
IPv6
3
UDP
4
TCP

• Supports
• Error control. For each segment
• Sequence numbers.
• Acknowledgment.
• Timeout
• Example stop and wait protocol. Very
  inefficient.
• Window based
• Flow Control
• Congestion Control

5
(No Transcript)
6
TCP
7
TCP 1 bit flags

• ACK when set the ack value is valid
• SYN, RST, FIN used for connection establishment
  and tear-down
• PUSH data should be passed to the upper layer
  immediately.
• URG there is urgent information in the data

8
Data streaming. MSS- maximum segment size
9
TCP flow control window
10
(No Transcript)
11
TCP is a connection-oriented protocol for
client-server communication
12
(No Transcript)
13
TCP congestion control

• Host centric, feedback-based resource allocation
  policy.
• The congestion control window is affected by the
  timing of the acknowledgments. A late or missing
  acknowledgment signals that the network is
  congested.

14
(No Transcript)
15
Other Internet protocols

• ICMP used by hosts and routers to exchange
  network layer information, e.g., error reporting
• RIP Routing Information Protocol
• OSPF Open Shortest Path First Protocol

16
Internet traffic

• TCP ? 90 95 of the Internet traffic
• 65-75 of TCP traffic is Web related
• 10 of TCP traffic is due to News
• 5 of TCP traffic is due to Email
• 5 of TCP traffic is due to FTP
• 1 of TCP traffic is due to Napster
• UDP ? 5 10 of the Internet traffic
• DNS
• Realaudio
• games

17
Scalability issues with VCs

• Assume a router with
• n input and n output lines of b Gbps
• average packet size l bytes
• determine the size of the routing table

18
(No Transcript)
19
Original Algorithm for Adaptive Retransmission

• Measure SampleRTT for each segment/ACK pair
• Compute weighted average of RTT
• EstimatedRTT a x EstimatedRTT (1- a) x
  SampleRTT
• where 0.8 lt a lt 0.9
• Set timeout based on EstimatedRTT
• TimeOut 2 x EstimatedRTT

20
Karn/Partridge Algorithm

• Do not sample RTT when re-transmitting
• Double timeout after each retransmission

21
Karn/Partridge Algorithm
22
Jacobson/Karels Algorithm

• New calculation for average RTT
• Diff SampleRTT - EstimatedRTT
• EstimatedRTT EstimatedRTT (d x
• Deviation Deviation d(Diff- Deviation)
• where d is a fraction between 0 and 1
• Consider variance when setting timeout value
• TimeOut m x EstimatedRTT f x Deviation
• where m 1 and f 4
• Notes
• algorithm only as good as granularity of clock
  (500 microseconds on Unix)
• accurate timeout mechanism important to
  congestion control (later)

23
Congestion Control Mechanisms

• The sender must perform retransmissions to
  compensate for lost packets due to buffer
  overflow.
• Unneeded retransmissions by the sender due to
  large delays causes a router to use link
  bandwidth to forward unneeded copies of a packet.
• When a packet is dropped along a path the
  capacity used used at each upstream routers to
  forward packets to the point where it was dropped
  was wasted.

24
Delay/Throughput Tradeoffs
25
(No Transcript)
26
Router with infinite buffer capacity
27
Fairness of TCP congestion mechanism
28
Flows and resource allocation

• Flow sequence of packets with a common
  characteristics
• A layer-N flow ? the common attribute a layer-N
  attribute
• All packets exchanged between two hosts ? network
  layer flow
• All packets exchanged between two processes ?
  transport layer flow

29
(No Transcript)
30
Min-max fair bandwidth allocation

• Goal fairness in a best-effort network.
• Consider
• Unidirectional flows
• Routers with infinite buffer space
• Link capacity is the only limiting factor.

31
Algorithm

• Start with an allocation of zero Mbps for each
  flow.
• Increment equally the allocation for each flow
  until one of the links of the network becomes
  saturated. Now all the flows passing through the
  saturated link get an equal fraction of the link
  capacity.
• Increment equally the allocation for each flow
  that does not pass through the first saturated
  link until a second link becomes saturated. Now
  all the flows passing through the saturated link
  get an equal fraction of the link capacity.
• Continue by incrementing equally the allocations
  of all flows that do not use a saturated link
  until all flows use at least one saturated link.

32
QoS in a datagram network?

• Packet Classification.
• Buffer acceptance algorithms.
• Explicit Congestion Notification.
• Flow measurements

33
Packet classification

• Identify the flow the packet belongs to.
• The edge routers may be able to do that.
• MPLS multi protocol label switch. Add an extra
  header in front of the IP header. Now a router
  decides the output link based upon the input link
  and the MPLS header.

34
Buffer acceptance algorithms

• Tail Drop.
• RED Random Early Detection
• RIO Random Early Detection with In and Out
  packet dropping strategies.

35
(No Transcript)
36
Explicit Congestion Notification (ECN)

• Routers could prevent congestion by informing the
  source of the packets when they become lightly
  congested, but before they start dropping
  packets.
• This strategy is called source quench.

37
Source quench

• A router sets a congestion notification flag in
  the IP header to inform the destination that
  signs of congestion are visible.
• The destination informs the source by setting a
  flag in the TCP header of segments carrying
  acknowledgments.

38
Problems with ECN

• (1) TCP must be modified to support the new flag.
• (2) Routers must be modified to distinguish
  between ECN-capable flows and those who do not
  support ECN.
• (3) IP must be modified to support the congestion
  notification flag.
• (4) TCP should allow the sender to confirm the
  congestion notification to the receiver, because
  acknowledgments could be lost.

39
Flow measurements

• How to choose the measurement interval to
  accommodate bursty traffic?
• Token bucket

40
The token bucket filter

• Characterized by (1) A token rate R, and (2)
  The depth of the bucket, B
• Basic idea the sender is allocated tokens at a
  given rate and can accumulate tokens in the
  bucket until the bucket is filled. To send a byte
  the sender must have a token. The maximum burst
  can be of size B because at most B token can be
  accumulated.

41
Example

• Flow A generates data at a constant rate of 1
  Mbps. Its filter will support a rate of 1 Mbps
  and a bucket depth of 1 byte,
• Flow B alternates between 0.5 and 2.0 Mbps. Its
  filter will support a rate of 1 Mbps and a bucket
  depth of 1 Mbps
• Note a single flow can be described by many
  token buckets.

42
Example
43
(No Transcript)
44
Token bucket

• L packet length
• C of tokens in the bucket
• --------------------------------------------------
  -
• if ( L lt C )
• accept the packet
• C C - L
• else
• drop the packet

45
A shaping buffer delays packets that do not
confirm to the traffic shape

• if ( L lt C )
• accept the packet
• C C - L
• else / the packet arrived early, delay it /
• while ( C lt L )
• wait
• transmit the packet
• C C - L

46
Packet Scheduling

• PS and GPS Processor Sharing Generalized
  Processor Sharing
• Round Robin, Weighted Round Robin
• Priority Scheduling
• Weighted Fair Queuing practical version of GPS.
  Transmits packets in the order of their finishing
  time.

47
Weighted queuing
48
RSVP- Resource Reservation Protocol

• Used to establish a path for a flow and reserve
  resources along the path.
• Requirements
• Accommodate faults soft state.
• Support unicast as well as multicast.
• PATH messages ? issued by sender includes TSpec
• RESV messages ? issued by the receiver includes
  RSpec

49
RSVP
50
RSVP message
51
RSVP multicast
52
Integrated Services

• Support fine-grain QoS for individual flows.
• Mechanisms
• Specification of flow requirements - Flowspecs
• Admission decisions
• Resource reservation and policing
• Policy enforcement

53
Flowspecs

• TSpec specify the traffic characteristics
• Rspec describe services required from network.

54
Admission decisions

• Two classes
• Guaranteed Services based upon token buckets
• Controlled Load approximates a best effort
  model in a lightly loaded network.

55
Integrated Service Router
56
Differentiated Services

• Two classes of traffic
• Regular
• Premium
• Edge routers mark the packets.
• Premium packets enjoy
• EF Expedited Forwarding
• AF Assured Forwarding