1 / 17

Data Transfer Case Study: TCP

Data Transfer Case Study: TCP. Go-back N ARQ 32-bit sequence # indicates byte number in stream transfers a byte stream, not fixed size user blocks full duplex (bi-rectional) data transfer sends upper level data "at its convenience" (RFC793!), trying to accumulate 512 bytes of data

saxon
Télécharger la présentation

Data Transfer Case Study: TCP

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. Data Transfer Case Study: TCP • Go-back N ARQ • 32-bit sequence # indicates byte number in stream • transfers a byte stream, not fixed size user blocks • full duplex (bi-rectional) data transfer • sends upper level data "at its convenience" (RFC793!), trying to accumulate 512 bytes of data • cumulative ACK: ACK(n) ack’s all bytes up through n • ACK for received A-to-B data piggybacked on B-to-A data packet • Internet checksum: add up data, take 1’s complement • covers both header and data

  2. TCP Packet Format 32 bits dest. port number source port number sequence number acknowledgment number u r g s y n a c k p s h r s t f i n head len window size unused ptr to urgent data Internet checksum options data

  3. Data Transfer: XTP • XTP: Xpress transfer protocol • designed for high-speed, high-performance networks • 32-bit sequence numbers with transition to 64-bit seq. numbers • 32-bit priority field for different priority data • user-selectable: reliable or unreliable data transfer • Go-Back-N ARQ but receiver can also indicate spans of packets received • sender only retransmits gaps

  4. Data Transfer: XTP • checksum: • form of two-dimensional parity • header and data checksummed separately • data checksumming can be disabled • data checksumming at end (send data while computing checksum (need only touch data once)

  5. Flow and Congestion Control sometimes sender shouldn’t send a ready packet: • receiver not ready (e.g., buffers full) • react to congestion • many unACK’ed packets may mean long end-end delays, congested networks • network itself may provide sender with congestion indication • avoid congestion: • sender transmits smoothly to avoid temporary network overloads

  6. Flow and Congestion Control flow control: speed and resource matching of sender and receiver • sender should not overwhelm receiver congestion control: action taken in response to network layer (and below) congestion • throttling sender is but one solution to congestion

  7. The flow control scenario

  8. Two approaches towards flow control Explicit flow control • receiver tells sender how much to send

  9. Explicit flow control (cont) Useful abstraction: sender maintains sliding window over sequence number, indicating what it can send • done in TCP and TP4 • congestion (as opposed to flow) control window may further restrict sender

  10. Flow Control in TCP Receiver explicitly advertises available buffer space to sender: • 16-bit "advertised window" specifies number of bytes (starting from ACKnum) receiver willing to receive • max window size is 64K • recent "scaling" option for larger windows

  11. <win=2048> <seq=1024,data> <seq=2048,data> <ACK=2049, win=1024> <seq=3072,data> Sender Receiver receiver advertises window size of 2048 bytes can send 1Kbytes, send 1K packet receiver delays ACK (no ACK generated) can send 1Kbytes, send 1K packet flow control window closed receiver generatess ACK ACK advertises window of 1024 bytes can send 1Kbytes, send 1K packet flow control window closed

  12. Implicit flow control • receipt of ACK’s triggers more sending • delayed ACK (for whatever reason) slows down sender • IBM virtual route pacing: • initially send window of N packets • ACK of first packet in this (and subsequent) windows of N allows sender to send N more • jumping", non-sliding window • max. number of unACK’ed packets? • decrease in traffic for explicit flow control

  13. Congestion Control Temporary demand for shared resources (links, processing, buffers) in network layer and below may exceed demand: • packets buffered until resources available • buffers full: packet lost (discarded) • which to discard? • lots of buffering: excessive delays network layer switching and routing input buffers output buffers

  14. Congestion Control: retransmission effects Ideal case: • every packet delivered successfully until subnet reaches capacity • beyond capacity: deliver packets at capacity rate In face of loss or long end-end delay, retransmissions can make things worse • inject more (not less) traffic into net • throwing fuel on fire!

  15. Congestion Control: retransmission effects

  16. Congestion Control (cont) Realistically: • as offered load increases, more packets lost, causing more retransmission, causing more traffic, causing more losses, … • at (b), each original packet sent four times on average • decreasing rate of transmission (e.g., a larger timeout value) increases overall throughput Moral: • when losses occur: backoff, don’t aggressively retransmit However: • social versus individual good: if most back off (and suffer) how to handle greedy sender who does not back off (and benefits)

  17. Three Basic Approaches towards Congestion Control End-end congestion control • sender-observed congestion (e.g., delays, losses) used to control sender • closed loop control Network-indicated congestion control • network layer provides feedback to sender Rate-based control • sender behavior fixed (bounded) over time • open-loop control Real-world protocols sometimes mix/combine these

More Related