Basic Architecture THE CONTINUOUS ROUND-ROBIN DISPATCHING SWITCH

CLOS-NETWORK SWITCHES 264 It is a challenge to find a cost-effective dispatching scheme that is able to achieve a high throughput in Clos-network switches, without allocating any buffers in the second stage to avoid the out-of-sequence problem and without expanding the internal bandwidth. w x A solution to the challenge was introduced in 14 , where a round-robin- based dispatching scheme, called the concurrent round-robin dispatching Ž . CRRD scheme, was proposed for a Clos-network switch. The basic idea of w x CRRD is to use the desynchronization effect 13 in the Clos-network switch. The desynchronization effect has been studied using simple scheduling w x Ž . w x algorithms as iSLIP 13, 11 and dual round-robin matching DRRM 1, 2 in an input-queued crossbar switch. CRRD provides high switch throughput without expanding the internal bandwidth, while the implementation is simple because only simple round-robin arbiters are employed. We showed that CRRD achieves 100 throughput under uniform traffic.

10.4.1 Basic Architecture

Figure 10.8 shows the CRRD switch. The terminology used in this section is as follows: IM Input module at the first stage. CM Central module at the second stage. OM Output module at the third stage. Ž . Ž . n Number of input ports output ports in each IM OM k Number of IMs or OMs. m Number of CMs. i IM number, where 0 F i F k y 1. j OM number, where 0 F j F k y 1. Ž . Ž . h Input port IP or output port OP number in each IMrOM, respectively, where 0 F h F n y 1. Ž . r Central module CM number, where 0 F r F m y 1. Ž . IM i ith IM. Ž . CM r r th CM. Ž . OM j jth OM. Ž . Ž . IP i, h hth input port at IM i . Ž . Ž . OP j, h hth output port at OM j . Ž . Ž . Ž . VOQ i, ® Virtual output queue VOQ at IM i that stores cells Ž . destined for OP j, h , where ® s hk q j and 0 F ® F nk y 1. Ž . Ž . Ž . G i, j VOQ group at IM i that consists of n VOQ i, j, h s. Ž . Ž . Ž . L i, r Output link at IM i that is connected to CM r i Ž . Ž . Ž . L r, j Output link at CM r that is connected to OM j . c THE CONTINUOUS ROUND-ROBIN DISPATCHING SWITCH 265 Fig. 10.8 CRRD switch with virtual output queues in the input modules. The first stage consists of k IMs, each of which is n = m. The second stage consists of m bufferless CMs, each of which is k = k. The third stage consists of k OMs, each of which is m = n. Ž . w x Ž . IM i has nk VOQs to eliminate HOL blocking 10 . VOQ i, ® stores cells Ž . Ž . Ž . that go from IM i to OP j, h at OM j , where ® s hk q j. A VOQ can receive, at most, n cells from n IPs in each cell time slot. The HOL cell in Ž . each VOQ can be selected for transmission across the switch through CM r in each time slot. This ensures that cells are transmitted from the same VOQ in sequence. Ž . Ž . Ž . IM i has m output links. An output link L i, r , is connected to CM r . i Ž . Ž . CM r has k output links, denoted as L r, j , and it is connected to k c Ž . OMs, denoted as OM j . Ž . Ž . OM j has n output ports, denoted as OP j, h , and has an output buffer. Each output buffer receives at most m cells in one time slot, and each output port of an OM forwards one cell in a FIFO manner to the output line.

10.4.2 Concurrent Round-Robin Dispatching Scheme