Performance of Stream Control Transmission Protocol (SCTP) Mohammed Atiquzzaman, Ph.D. School of Computer Science University of Oklahoma. Email:
[email protected] Web: www.cs.ou.edu/~atiq
Tohoku University, Sendai, Japan. Aug 6, 2002. Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
1
Introduction ! TCP is the main transport protocol in the Internet protocol suite ! Original TCP performed poorly in satellite networks " errors " long propagation delay.
! Many schemes for enhancing TCP for satellite networks. ! IETF developing Stream Control Transmission Protocol (SCTP) for PSTN signaling over IP.
Open question: How good is multistreaming?
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
2
Objectives ! Evaluate the performance of SCTP multistreaming. ! Suitability of SCTP for wireless networks ! Performance comparison of SCTP and TCP.
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
3
Outline ! SCTP " Multistreaming " Multihoming
! Performance of multistreaming
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
4
Stream Control Transmission Protocol ! SCTP (RFC 2960) being developed by IETF as a transport protocol for PSTN signaling. " Reliable: retransmission of lost packets, ack of packets. " Non-duplicated service: uses sequence numbers. " In-order delivery: re-sequencing at the destination.
! Transport layer protocol which operates on top of an unreliable connectionless network layer such as IP. " Transparent to IPv4 or IPv6
! Key features: " Multistreaming – multiple streams per association " Multihoming – multiple IP addresses per host
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
5
SCTP in the protocol stack
Upper layer applications TCP, UDP, SCTP IP Link Layer Physical Layer
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
6
Multistreaming ! SCTP accomplishes multistreaming by creating independence between " data transmission (uses Transport Sequence Number) " data delivery (uses Stream Sequence Number)
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
7
Multihoming ! Supports multiple IP addresses in an association. ! Requires multiple Network Interface Cards – already quite common in laptops !!
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
8
SCTP Packets
Chunk Type •Payload •SACK, etc.
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
9
Chunk Type: Payload
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type = 0 | Reserved|U|B|E| Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | TSN | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Stream Identifier S | Stream Sequence Number n | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Payload Protocol Identifier | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ / / / User Data (seq n of Stream S) / / / +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
10
Chunk Type: SACK
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type = 3 |Chunk Flags | Chunk Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Cumulative TSN Ack | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Advertised Receiver Window Credit (a_rwnd) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Number of Gap Ack Blocks = N | Number of Duplicate TSNs = X | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Gap Ack Block #1 Start | Gap Ack Block #1 End | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ / / / ... / / / +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Gap Ack Block #N Start | Gap Ack Block #N End | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Duplicate TSN 1 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ / / / ... / / / +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Duplicate TSN X | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
11
SCTP Congestion Control ! SCTP congestion control is similar to TCP congestion control. " Enables seamless introduction of SCTP into IP networks.
! SCTP is rate adaptive similar to TCP. # Slow Start, Congestion Avoidance, Fast Retransmit # Fast Recovery is implemented, but in a slightly different way than TCP.
! Differences with TCP " Number of bytes acknowledged is used to increase cwnd. " SACK is mandatory # Unlimited number of Gap Ack Blocks in SACK " No explicit fast recovery phase
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
12
Performance/Advantages of SCTP Multistreaming
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
13
ns-2 Simulation Setup
! Ftp traffic. ! Packets of fixed length (one MTU). ! Upper layer at destination is always ready to accept data. ! Association consists of a number of streams
Mohammed Atiquzzaman, University of Oklahoma, USA.
Link Delay: L1+L2 = 260 msec Receiver buffer size = B
Presentation at Tohoku University, Japan Aug 6, 2002.
14
Performance Metrics ! Goodput ! Optimal receiver buffer size " as a function of packet error probability (e).
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
15
Packet Plot: No loss
s = 4, e = 0, B = 15K
! No packet loss $ no blocking at receiver. ! cwnd increases until B. ! Goodput limited to B/MTU packets per RTT.
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
16
Goodput: No loss
s = 4, e = 0
! Since goodput is limited to B/MTU packets every RTT; it increases linearly with B.
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
17
Packet plot: Congestion Control limited ! Poor goodput when receiver buffer is not a constraint
s = 4, e = 0.01, B=35K
" Long delays in Retx while waiting for DupAcks
+ " drop in cwnd due to Retx
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
18
cwnd and a_rwnd with errors: Receiver buffer limited
s = 4, e = 0.01, B = 15K
! B=15K $ throughput constrained by receiver buffer
Mohammed Atiquzzaman, University of Oklahoma, USA.
cwnd is restricted to 15K
a_rwnd frequently drops below 1 MTU
Presentation at Tohoku University, Japan Aug 6, 2002.
19
cwnd and a_rwnd with errors: Congestion control limited
s = 4, e = 0.01, B = 35K
! B=35K $ throughput constrained by congestion control of SCTP
a_rwnd usually does not drop below 1 MTU
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
20
Goodput with errors s = 4, e > 0
Congestion control limited: !Goodput is limited by the congestion control of SCTP. ! Goodput can only be increased by lowering the error rate
Receiver buffer limited: ! Goodput increases as B increases when the goodput is constrained by B (a_rwnd frequently drops below 1 MTU) Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
21
Goodput vs. Buffer size: One and four streams
! For small B Mutlistreaming results in “less” HOL blocking " goodput of 4streams is higher than 1-stream.
! For large B Goodput is limited by congestion control
Multistreaming increases goodput for small receiver buffers. Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
22
Advantage of Multistreaming: High Throughput
B = 15K
! “Small” Buffer size of 15K shows advantage of multistreaming. Mohammed Atiquzzaman, University of Oklahoma, USA.
HOL blocking is eliminated as evidenced by the fact that a_rwnd is, not a limiting Presentation at Tohoku University, Japan factor Aug 6, 2002.
23
Optimal Buffer Size
Optimal Receiver Buffer Size Buffer size beyond which a_rwnd never falls below 1 MTU
Multistreaming reduces receiver buffer requirements. Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
24
Conclusions ! Performance of SCTP multistreaming has been studied ! Multistreaming increases transport level performance for small receiver buffer size. " Wireless handheld devices, ex. Mobile and satellite networks.
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
25
Further information ! Acknowledgements " National Aeronatics and Space Administration (NASA)
! Further Information Dr. Mohammed Atiquzzaman
[email protected], (405) 325 8077 ! These slides are available at www.cs.ou.edu/~atiq
Mohammed Atiquzzaman, University of Oklahoma, USA.
Presentation at Tohoku University, Japan Aug 6, 2002.
26
