iSLIP Switch Scheduler Ali Mohammad Zareh Bidoki April 2002

Table of Contents The place Buffer in Crossbar Switches Example of Fabrics PIM iSLIP (in CISCO 12000,5Gb/s router and Tiny Tera 0.5 Tb/s) RRM WFA PP_VOQ Multicasting A 2.5Tb/s Router

The place of Buffer in Crossbar Output Buffer Shared Buffer Input buffer

Interconnects Two basic techniques Input Queueing Output Queueing Usually a non-blocking switch fabric (e.g. crossbar) Usually a fast bus

Interconnects Input Queueing with Crossbar configuration Data In Data Out Arbiter Memory b/w = 2R

Input Queueing Head of Line Blocking Delay Load 58.6% 100%

Head of Line Blocking

Virtual output Queuing Crossbar Switch fabric To port 1 Input queues To port 2 Port n queue To port n Port 2 queue Port 1 queue Input port 1 Queue scheduler

Input Queueing Virtual output queues

Input Queueing Virtual Output Queues Delay Load 100%

Which is better? Virtual output Queue (input queue). Ideal Output queue.

Input Queueing Virtual output queues Arbiter Complex!

VOQ Arbiter Input memory management

Problem Definition ( bipartite)

Maximum or Maximal matching

Maximum matching Maximizes instantaneous throughput Starvation Time complexity is very high in Hardware (o(n 3 )) Maximal matching Can’t add any connection on the current match without alert existing connections More practical (e.g. WFA, PIM, iSLIP, DRR,RRM)

Matching Algorithms Each algo. is evaluated by four parameters: 1.Latency(Throughput). 2.Starvation free. 3.Fast. 4.Implementation. 3. iSLIP – Iterative Serial-Line IP(base on PIM and RRM) 2. RRM – Round-Robin Matching 1. PIM - Parallel Iterative Matching We will discuss three different matching algo.:

When no new matching can be found, the algorithm stops. 3.Accept-If an input receives a grant, it accepts one by selecting an output randomly among those that granted to this output.. 2.Grant-If an unmatched output receives any requests, it grants to one by randomly selecting a request uniformly over all requests. 1.Request-Each unmatched input sends a request to every output for which it has a queued cell. PIM - Parallel Iterative Matching The basic matching algorithm. Each iteration of the algorithm follows these three steps:

PIM Each iteration will eliminate at least ¾ of the remaining connections Converge in O(logN) iterations No input queue is starved if service No memory or state is used At the beginning of each cell time, the match begins over, independently of the matches that were made in previous cell times PIM does not perform well for a single iteration: it limits the throughput to approximately 63%, only slightly higher than for a FIFO switch. This is because the probability that an input will remain ungranted is (N-1/N) N, hence as N increases, the throughput tends to.63% (1-(1/e)) Implementation is hard in Hardware

RRM – Round-Robin Matching The pointer g i to the highest priority element of the round-robin schedule is incremented (modulo N) to one location beyond the granted input. 2.Grant-If an output receives any requests, it chooses the one that appears next in a fixed, round-robin schedule starting from the highest priority element. The output notifies each input whether or not its request was granted. 1.Request-Each unmatched input sends a request to every output for which it has a queued cell. g2g2 g4g4 g1g1 a1a1 a3a3 a4a

RRM – Round-Robin Matching The pointer a i to the highest priority element of the round-robin schedule is incremented (modulo N) to one location beyond the accepted output. 3.Accept-If an input receives a grant, it accepts the one that appears next in a fixed, round-robin schedule starting from the highest priority element. The pointer g i to the highest priority element of the round-robin schedule is incremented (modulo N) to one location beyond the granted input. 2.Grant-If an output receives any requests, it chooses the one that appears next in a fixed, round-robin schedule starting from the highest priority element. The output notifies each input whether or not its request was granted. 1.Request-Each unmatched input sends a request to every output for which it has a queued cell. a1a1 a3a3 a4a g2g2 g4g4 g1g1

RRM – Round-Robin Matching The pointer a i to the highest priority element of the round-robin schedule is incremented (modulo N) to one location beyond the accepted output. 3.Accept-If an input receives a grant, it accepts the one that appears next in a fixed, round-robin schedule starting from the highest priority element. The pointer g i to the highest priority element of the round-robin schedule is incremented (modulo N) to one location beyond the granted input. 2.Grant-If an output receives any requests, it chooses the one that appears next in a fixed, round-robin schedule starting from the highest priority element. The output notifies each input whether or not its request was granted. 1.Request-Each unmatched input sends a request to every output for which it has a queued cell. a1a1 a3a3 a4a g2g2 g4g4 g1g1

RRM – Round-Robin Matching g2g2 g3g3 g1g1 a1a1 a2a2 a3a First cycle The RRM is not starvation free: In the following example, we assume there are always cells waiting to be transferred. The destination is always the same.

RRM – Round-Robin Matching g2g2 g3g3 g1g1 a1a1 a2a2 a3a First cycle The RRM is not starvation free: In the following example, we assume there are always cells waiting to be transferred. The destination is always the same.

RRM – Round-Robin Matching g2g2 g3g3 g1g1 a1a1 a2a2 a3a First cycle The RRM is not starvation free: In the following example, we assume there are always cells waiting to be transferred. The destination is always the same.

RRM – Round-Robin Matching g2g2 g3g3 g1g1 a1a1 a2a2 a3a First cycle The RRM is not starvation free: In the following example, we assume there are always cells waiting to be transferred. The destination is always the same.

RRM – Round-Robin Matching g2g2 g3g3 g1g1 a1a1 a2a2 a3a First cycle The RRM is not starvation free: In the following example, we assume there are always cells waiting to be transferred. The destination is always the same.

RRM – Round-Robin Matching g2g2 g3g3 g1g1 a1a1 a2a2 a3a Second cycle The RRM is not starvation free: In the following example, we assume there are always cells waiting to be transferred. The destination is always the same.

RRM – Round-Robin Matching g2g2 g3g3 g1g1 a1a1 a2a2 a3a Second cycle The RRM is not starvation free: In the following example, we assume there are always cells waiting to be transferred. The destination is always the same.

RRM – Round-Robin Matching The RRM is not starvation free: In the following example, we assume there are always cells waiting to be transferred. The destination is always the same. g2g2 g3g3 g1g1 a1a1 a2a2 a3a Second cycle At this point the sequence of the events will repeat itself: Outputs 1 and 3 will always grant input 1, while output 2 will always grant input 1 at the first iteration of the first cycle, but input 1 will select output 1 indefinitely, leaving output 2 to grant either input 2 or input 3. Thus the cell from input 1 to output 2 will never be granted. In order to solve this starvation the iSlip algorithm was developed.

RRM RRM overcomes two problem Complexity Unfairness the round-robin arbiters are much simpler and can perform faster than random arbiters. The rotating priority aids the algorithm in assigning bandwidth equally and more fairly among requesting connections. Its throughput is about 63%

2x2 switch with RRM algorithm under heavy load. synchronization of output arbiters leads to a throughput of just 50%.

Performance

Synchronization

iSLIP – Iterative Serial-Line IP 2.Grant-If an output receives any requests, it chooses the one that appears next in a fixed, round-robin schedule starting from the highest priority element. The output notifies each input whether or not its request was granted. g2g2 g4g4 g1g1 a1a1 a3a3 a4a The pointer g i to the highest priority element of the round-robin schedule is incremented (modulo N) to one location beyond the granted input if and only if the grant is accepted in Step 3 of the first iteration.

iSLIP – Iterative Serial-Line IP The pointer g i to the highest priority element of the round-robin schedule is incremented (modulo N) to one location beyond the granted input if and only if the grant is accepted in Step 3 of the first iteration. 2.Grant-If an output receives any requests, it chooses the one that appears next in a fixed, round-robin schedule starting from the highest priority element. The output notifies each input whether or not its request was granted. g2g2 g4g4 g1g a1a1 a3a3 a4a4

iSLIP properties Property 1. Lowest priority is given to the most recently made connection. If input i successfully connects to output j, both a i and g j are updated and the connection from input i to output j becomes the lowest priority connection in the next cell time. Property 2. No connection is starved. This is because an input will continue to request an output until it is successful. The output will serve at most other inputs first, waiting at most N cell times to be accepted by each input. Therefore, a requesting input is always served in less than N 2 cell times. Property 3. Under heavy load, all queues with a common output have the same throughput. This is a consequence of Property 2: the output pointer moves to each requesting input in a fixed order, thus pr-viding each with the same throughput.

iSLIP properties Simple to implement in hardware Starvation free Its throughput is about 100% It is fair As the load increases, the number of synchronized arbiters decreases (see Figure), leading to a large sized match. Under uniform 100% offered load the iSLIP arbiters adapt to a time-division multiplexing scheme. It converge in O(1)

Bursty Arrivals

Burstiness Reduction Results indicate that iSLIP reduces the average burst length, and will tend to be more burst-reducing as the offered load increases. This is because the probability of switching between multiple connections increases as the utilization increases. As the load increases, the contention increases and bursts are interleaved at the output. In fact, if the offered load exceeds approximately 70%, the average burst length drops to exactly one cell.

Burstiness Reduction

Multiple Iteration The pointer g i to the highest priority element of the round-robin schedule is incremented (modulo N) to one location beyond the granted input if and only if the grant is accepted in Step 3 of the first iteration. Note that pointers g i and a i are only updated for matches found in the first iteration. It converge in O(logN)

Multiple Iteration

All with 4 iterations

Implementation

Implementation(2N arbiters)

Implementation(N arbiters) Each arbiter is used for both input and output arbitration. In this case, each arbiter contains two registers to hold pointers g i and a i.

Implementation

Priority in iSLIP

Why iSLIP is good for high speed? input buffers are separated Separated scheduler for each input and output Each work independently

Multicasting Fanout splitting: higher throughput, but not as simple Non-fanout splitting: Easy, but low throughput

Multicasting ( ESLIP: Combining Unicast and Multicast-use in CISCO )

IP packet in iSLIP switch ( 2N 2 Queue) Arbiter

Linecard LCSLCS LCSLCS 1: Req LCS Ingress Flow control(2.5Tb/s) 3: Data Switch Scheduler Switch Scheduler 2: Grant/credit Seq num Switch Fabric Switch Fabric Switch Port Req Grant

LCS Over Optical Fiber 10Gb/s Linecards 10Gb/s Linecard LCSLCS Switch Scheduler Switch Scheduler Switch Fabric Switch Fabric 10Gb/s Switch Port LCSLCS 12 multimode fibers 2.5Gb/s LVDS GENET Quad Serdes

2.56Tb/s IP router LCS 1000ft/300m Port #256 Port #1 2.56Tb/s switch core Linecards

Port Processor optics LCS Protocol optics Port Processor optics LCS Protocol optics Crossbar Switch core architecture Port #1 Scheduler RequestGrant/CreditCell Data Port #256