Qos Guarantees
Qos Guarantees
introduction
call
admission traffic specification link-level scheduling call setup protocol reading: Tannenbaum, 393-395, 458-471
Ch 6 in Ross/Kurose
Motivation
Certain applications require minimum level of network performance:
Internet
telephone, teleconferencing: delays > 500ms impair human interaction session guaranteed QoS or is blocked (denied admission to network) starting to look like telephone net!
Fundamental mismatch between QoS and packet switching packet switching: statistically share resources in hope that sessions' peak demands don't coincide
100% certain guarantees require accounting for worst case behavior (no matter how unlikely)
admitting/denying
Next generation Internet service classes: guaranteed service: "provides firm (mathematically provable) bounds on end-to-end datagram queuing delays. This service makes it possible to provide a service that guarantees both delay and bandwidth." controlled load: "a QoS closely approximating the QoS that same flow would receive from an unloaded network element, but uses capacity (admission) control to assure that this service is received even when the network element is overloaded." best effort: current service model
Question: what new concepts, mechanisms needed to insure that packets from source see a given QoS?
Questions to be answered:
how
much traffic will be injected by call into net, how should that traffic be described (is rate (pkts/sec) enough)? what if call sends more traffic than it claimed it would? QoS requirement is end-to-end, how to break into per-hop requirements? what resources will be needed (bandwidth, buffering) to meet QoS requirement? how to reserve resources for call at each hop: call setup protocol current networks: routers play no role in call admission
leaky bucket proposal: traffic entering net filtered by leaky bucket regulator: b: maximum burst size r: average rate amount of traffic entering over any interval of length t, less than b + rt
pkts from entering net (shaping) drop pkts that arrive without tokens (policing function) let all pkts pass thru, mark pkts: those with tokens, those without network drops pkts without tokens in time of congestion (marking)
Round-robin service: assume fixed length pkts, N sessions session i gets to send 1 pkt each "turn" if session i has no pkt, i+1 gets chance
of session 1 traffic < r_1*t + b_1 burst of size b_1 arrives, empties bucket, enters queue 1 time until last packet served (maximum pkt delay) is b_1 / (f_1 C) queue length decreases from b_1 (assuming r_1 < f_1/sum(all j, f_j) * C) Analysis can be extended to networks of multiplexors
Reserving Resources
call
setup protocol needed to perform call admission, reserve resources at each router on end-end path
Q.2931:
ATM call setup protocol sender initiates call setup by passing Call Setup Message across UNI boundary (i.e., into network) network immediately returns Call Proceeding indication back network must:
choose path (routing) allocate resources along path (note: QoS and routing coupling)
network
large numbers of heterogeneous receivers heterogeneity in available bandwidth to receiver heterogeneity in receiver QoS demands (differing delay requirements)
scalability heterogeneity
Soft state: call state (reservations) in routers need not be explicitly deleted
RSVP: Multicast
Multiple receivers:
may
have different QoS requirements resource reservations merged as RESV travels upstream resources must be allocated to satisfy strictest demands of downstream receivers e.g.: receiver 1 first reserves resources for 100 ms max delay if receiver 2 QoS is 200 ms max delay, no new resources at R2, R1 if receiver 2 QoS is 50 ms, more resources needed at R2, R1
Can handle multiple senders as well: different "styles" of resource reservation e.g., reserve enough resources in case all senders simultaneous resource enough resources for two simultaneous senders can dynamically determine which of N streams is forwarded downstream (switching within the network)