Congestion Control

• TCP Vegas
• other ways of detecting congestion
• addressing congestion
• router intervention
• Internet Explicit Congestion Notification

TCP Vegas

• Reno detects congestion after it happens
• Reno also causes congestion by increasing the window until congestion occurs
• early congestion detection: as queues get filled in the router, packets take longer, so the RTT increases
• when RTT gets bigger, we can slow our sending
• when RTT gets back to minimum, we can increase our sending
• not standard, but tested to work well

Detecting and Addressing Congestion

• detecting congestion:
  • queues get longer
  • RTT gets bigger
  • data / RTT ( power) starts to drop as you try to send more
• addressing congestion:
  • additive increase/multiplicative decrease (needed for stability if congestion is occurring)
  • additive increase/additive decrease (TCP Vegas) -- works as long as congestion can be avoided
  • setting flow rate
  • bandwidth reservation

Router Intervention to Avoid or React to Congestion

• Random Early Discard -- causes TCP to back off
• information feed-forward -- the receiver must then return congestion information to the sender (see Internet ECN, below)
• information feedback -- requires route back to sender, does not work in Internet (except source quench ICMP, which is deprecated)
• communication time from router to sender may be insufficient if sender is sending lots of stuff. Also, stability issues -- all senders could increase their sending rate at the same time
• credits: can only send as much as we have in the "bank", automatically (but not immediately) replenished

Internet Explicit Congestion Notification

• ECN, explicit congestion notification, RFC 3168.
• in ECN, two of the bits of the Type of Service (ToS) field are used to indicate (a) whether congestion notification is requested (ECT), and (b) whether the packet experienced congestion (CE).
• TCP uses two new bits: ECE (ECN-Echo, to report that a packet was received with the CE bit set -- bit before URG), and CWR (Congestion Window Reduced, bit before ECE), to indicate that the ECE bit was received.
• compatible with hosts and routers that don't do ECN
• typical usage of ECN:
  • senders can set ECT
  • routers can change ECT to CE to record that congestion was experienced, perhaps instead of dropping a packet
  • transport layer is informed of CE, sends an ECE
  • receiver of ECE reduces congestion window, sends CWR

TCP Vegas

• packet scheduling
• FIFO queueing
• fair queueing

Packet Scheduling

• routers have a queue of outgoing packets for each interface
• any number of threads can place packets in the queue
• a single thread removes and sends the front packet in the queue whenever the previous packet has been sent
• there can be many strategies for placing packets in the queue!

FIFO Queuing

• each packet is placed at the end of the queue
• packets (that take the same route) are never reordered
• delay is proportional to queue size
• works reasonably well in Internet, with TCP congestion control
• if all but one sender do congestion control, and one does not, the one that doesn't (IP telephony, multicasting) might grab much of the bandwidth

Fairness

• "everyone" gets the same treatment
• hard to do in a distributed system:
• local fairness (every flow gets the same treatment on this router) discriminates against flows that cross more routers (parking garage problem)
• global fairness requires global co-ordination, so local fairness is often the best we can do

Fair Queueing

• one FIFO queue for each flow
• packets are taken in round-robin order from each queue that has them
• problem: large packets counted the same as small packets
• logically, we want to send one bit from each flow in round-robin order

Fair Queueing with Different size packets

• virtual clock ticks once for each bit sent from all the queues (so more active queues means slower virtual clock)
• virtual finish time for a packet is start time plus the size of the packet
• virtual start time is the largest of: finish time for the previous packet in the queue (a computed quantity), or arrival time
• select and transmit the packet with the lowest virtual finish time

TCP Review

• state management: connection setup and teardown, Transmission Control Block (TCB)
• reliable transmission via sequence numbers, acknowledgements
• flow control to avoid overwhelming receiver:
  • challenges obtaining both reliability and performance
  • acknowledgements are not acknowledged, but crucial information carried in the acknowledgements (e.g. window size)
• congestion control to avoid overwhelming network (or to slow down when we do) requires adaptive timer
• "good enough" (TCP Reno) may be preferable to "better" (TCP Vegas)

Our network so far

• TCP and IP
• reliable byte-stream and packet transmission among applications
• only application so far: DNS
• only Data Link layer so far: SLIP
• only Physical layer so far: serial lines, which are limited in speed (about 100,000 b/s max), distance (typically within a building), and number of hosts (at most two hosts per serial line)