1 / 15

Structured Streams: a New Transport Abstraction

Structured Streams: a New Transport Abstraction. Sigcomm 2007 By Bryan Ford Presented by Kyungbaek Kim ( kyungbak@uci.edu ). Introduction. Streams Extended Lifetime In-order delivery Datagrams Ephemeral Lifetime Independent delivery

sirvat
Download Presentation

Structured Streams: a New Transport Abstraction

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. Structured Streams: a New Transport Abstraction Sigcomm 2007 By Bryan Ford Presented by Kyungbaek Kim ( kyungbak@uci.edu )

  2. Introduction • Streams • Extended Lifetime • In-order delivery • Datagrams • Ephemeral Lifetime • Independent delivery • An application with a mixture of many small and a few large transactions faces awkward tradeoffs • Efficiency of UDP for small transactions • Reliability of TCP for the occasional large ones

  3. Problems of TCP streams • High network cost • Setup cost : 3way handshake and slow start per stream, • Shutdown cost : 4 minute TIME WAIT period • Head-of-line blocking • One loss delay everything behind • TCP unusable for real-time video/voice conferencing • HTTP/1.1 made web browsers slower! [Nielsen/W3C] • Unnecessary serialization • no parallelism between application data units (ADUs)

  4. Goals • Make streams cheap • Let application use one stream per ADU, efficiently • Make streams independent • Preserve natural parallelism between ADUs • Make streams easy to manage • Don't have to bind, pass IP address & port number, separately authenticate each new stream

  5. SST organization

  6. Streams, channels and packets

  7. SST Packet header

  8. Channel Protocol : Sequencing Every transmission gets new packet sequence # – Including acks, retransmissions [DCCP]

  9. Channel Protocol : Acknowledgement • All acknowledgments are selective [DCCP] • Each packet acknowledges a sequence range • Successive ACKs usually overlap : redundancy against lost ACKs • No variable-length SACK headers needed : all info in fixed header ASN = 1, AckCt = 1 ASN = 1, AckCt = 2 ASN = 1, AckCt = 3 ASN = 5, AckCt = 1 ASN = 5, AckCt = 2 ASN = 5, AckCt = 3

  10. Stream Protocol : Creating Streams • All segments during 1st round-trip carry “create” info • (special segment type, parent & child stream IDs) • Child borrows from parent stream's receive window • (“create” packets belong to parent stream for flow control) Init packet

  11. Stream Protocol : Data Transfer • Regular data transfer (after 1st round-trip): • 32-bit wraparound byte sequence numbers (BSNs) • Unlimited stream lifetime Data packet

  12. Stream Protocols : Best-effort Datagrams • When datagram is small: • Stateless best-effort delivery optimization • (avoids need to assign stream identifier to child) • When datagram is large: • Fall back to delivery using regular child stream Datagram packet

  13. Transaction Microbenchmark

  14. Web Browsing Workloads • Performance of transactional HTTP/1.0 on SST: • Much faster than HTTP/1.0 on TCP • Faster than persistent HTTP/1.1 on TCP • As fast as pipelined HTTP/1.1 on TCP

  15. Critique • Channel ID can limit the number of connected hosts • Channel ID is limited to 8 bits • It is not proper to any server program • Why is the lifetime of channel limited? • 64 bit sequence number space limit the lifetime • For security? • Unclear channel migration mechanism • What is the threshold of the best-effort datagram? • The stream layer checks that the datagram is small enough to ensure a reasonable chance of successful delivery

More Related