Peng Yi

Load-Balanced Multipath Self-Routing Switching Structure by Concentrators

A novel two stage load-balanced multipath self-routing switch structure is introduced in this paper. Both stages use a multipath self-routing fabric. With simple algorithms and small buffers, the first stage fabric transforms the incoming traffic into uniform and the second stage fabric forwards the data in a self-routing manner to their final destinations. Compared with other similar structures, this structure outstands with no queuing delay and zero jitter, its component complexity and propagation delay are significantly reduced. Mathematical analysis and simulations show this structure can achieve 100% throughput under admissible traffic pattern, which is a common presumption for incoming traffic. For statistically admissible traffic, by stacking up a few copies of this structure, it is suitable to support QoS application for building super large scale switching fabric in next generation network(NGN).

Multi-path Self-routing Switching Structure; by Interconnection of Multistage Sorting Concentrators

Several nonblocking packet switching architectures have been proposed for broadband network, such as shared bus, shared memory, crossbar matrix with combined input and output queuing, etc. Their topological demerit, such as bandwidth bottleneck and insufficient processing ability to schedule I/O matching, greatly limits their ability for large scale switching routers. This paper proposes and models a novel multi- path self-routing switching fabric by merging the bitonic sorters with the multistage interconnection network. This structure possesses the properties of complete distributing and self-routing, free of I/O matching scheduling algorithm, no internal buffer, no buffered delay and jitter, modeled with algebraic permuting group, as well as high modularity and recursive scalability. Mathematical analysis and simulations show this structure is suitable for building super large scale switching fabric to support QoS applications.

How to Make Network Nodes Adaptive

The current Internet lacks in adaptability fueling the great interest in defining a new network architecture that can meet the needs of a future Internet. One of the prevailing trends in this context is re-splitting the network function into fine granularity build blocks to breaking through the network ossification and realizing the network functional composition for enhancing network adaptability. In our work, we propose a novel adaptive architecture, i.e., ReNet, differing from existing solutions, which re-abstracts the current protocol stack at special location, and decompose the network and transport layers into the atomic capacities which is finer functional building blocks, and open the network core for adding new building blocks. Then, a nonlinear integer optimal problem is formulated for the composition of atomic capacities driven by users requests, with the proposed algorithm to reach appropriate tradeoff between optimal solution and computation cost. Numerical results demonstrate our algorithm can combine the atomic capacities in a feasible scale between cost and optimization. Finally, we give a proof-of-concept paradigm.

Integrated uni- and multicast traffic scheduling in buffered crossbar switches

The CICQ switches have inherent advantages in supporting multicast traffic, and scheduling policies can be simplified to a great extent when distributed scheduling is applied. However, simulation results indicate that the increment of multicast will block unicast, and finally reduce the performance of the entire switch. To overcome this deficiency, the paper proposes a new switch called MCICQ (multicast supporting CICQ) which supports the uni- and multicast hybrid scheduling, and scheduling policies MFDR and MFS which support the parallel distributed uni- and multicast are also proposed. With the fanout splitting mechanism, both MFDR and MFS can achieve good performance without speedup, and their complexities are only O(1). MFS scheduling policy can also provide different bandwidth guarantees for uni- and multicast traffic. Simulation results under SPES show that both MFDR and MFS can achieve good performance.

PIFO Output Queued Switch Emulation by a One-cell-Crosspoint Buffered Crossbar Switch

It is well known that the buffered crossbar has simpler scheduling algorithms than an unbuffered crossbar. Buffered crossbar can be pipelined to run at a high speed, making it appealing for high performance switches and routers. Recent researches indicate that a buffered crossbar with modest speedup can exactly emulate an output queued (OQ) switch. As for per flow/priority guarantee, additional speedup and storage is required to avoid crosspoint blocking, preventing the use of the buffered crossbar for lager-scale devices. This paper introduces a novel mechanism to solve crosspoint blocking and a simple architecture to provide per flow/priority guarantee. Based on a simple scheduling scheme, named modified group-by-first-in-first-out-group-lowest time-to-leave (MGBFG-LTTL), we sufficiently prove that a one-cell-crosspoint buffered crossbar with input virtual priority output queues, VPOQ/CB-1, switch with two times speedup can exactly emulate a push-in-first-out (PIFO) OQ switch. Our scheme has less hardware requirements and provides a simple path to scale crossbar based routers

PMUF: A High-Performance Scheduling Algorithm for DiffServ Classes

The differentiated services (DiffServ) model relies on the Per Hop Behavior (PHB) in each network device to treat packets differently based on packet headers’ DSCP (DiffServ Code Point). For each DiffServ compliant router, the scheduling algorithm is critical in implementing PHBs, according to which packets are forwarded. In this paper, we propose the parallel maximum urgency first (PMUF) scheduling algorithm for combined input-crosspoint queued (CICQ) switches to support DiffServ classes. The proposed PMUF algorithm features in a distributed scheduling scheme which can be implemented on each input and each output independently and in parallel. It adopts a two-stage flow control mechanism based on periodic statistic to provide minimum bandwidth guarantees for EF and AF traffic, and uses a priority scheduling mechanism to provide lower delay for EF traffic. The time complexity of PMUF is only O(log N), hence is practical and scalable for high speed application. Simulation results show that PMUF provides minimum bandwidth guarantees for EF and AF traffic and fair bandwidth allocation for BE traffic. PMUF also exhibits better delay performance than existing maximal matching based DiffServ scheduling schemes especially under non-uniform traffic.

A Fair Service and Dynamic Round Robin scheduling scheme for CICQ switches

The limitations in complexities and extensibilities of CICQ switchespsila scheduling policies are first analyzed. Then, based on this analysis, the guidelines for designing high extensible scheduling policies and the concept of virtual channel are proposed. Based on the guidelines and virtual channel, it comes up with a dynamic round robin scheduling algorithm-FDR (fair service and dynamic round robin), which is simple, high efficiency and fair service. FDR is based on round robin mechanism and its complexity is O(1). It allots the scheduling share for each virtual channel according to its current states. Thus, FDR has good dynamic and real-time performance, and it can adapt to unbalanced traffic load network environment. Simulation results under SPES show that FDR performs good delay, throughput and anti-burst performance, which can be applied in high performance routing and switching devices.

Towards Adaptive Network Nodes via Service Chain Construction

Network functional combination is a promising direction in enhancing Internet adaptability. It decomposes the current layered network into fine-grained building blocks and combines them on demand. However, what legacy functions should be decomposed and how to combine them in an optimal way are unclear. We propose a novel adaptive architecture called reconstructive network architecture (RECON) based on the principles of the Complex Adaptive System. This study has three main contributions. First, RECON decomposes functions of the protocol stack at layers 3 and 4 into fine-grained building blocks, called atomic capabilities to open the network core functions unlike existing solutions. Second, RECON can customize different service chains on demand by combining atomic capabilities in an optimal way. We formulate the atomic capability combination into a nonlinear integer optimization problem with the proposed algorithm to reach an appropriate tradeoff between the optimal solution and computation cost. Finally, we implement a proof-of-concept for RECON in the network node. Results are corroborated by several numerical simulations.

Sizing Buffers for Pipelined Forwarding Engine

Packet buffers in routers constitute a central element of packet networks. Selecting appropriate buffer size is an important and open research problem. This paper aims to size buffers for forwarding engine known as one of the main parts of a router. First, a high-speed pipeline which is designed for forwarding engine is presented, and its memory analysis model is also given, then, the memory requirement of the forwarding engine is analyzed under two conditions: the forwarding rate is not less than and less than the input rate. Our analysis results and experiment both show that, the proposed forwarding pipeline is of high performance, and just one pipeline can deal with the data transfer rate of 20 Gb/s or even higher; the pipelined forwarding engine only need to buffer a several packets, then the loss rate will be an acceptable value or even zero, further increasing the buffer size will have little effect on reducing the loss rate

A New Practical Scheduling Algorithm for Buffered Crossbar Switches

Recently, buffered-crossbar (CICQ) switch is becoming more and more attractive to high performance router builders than bufferless schemes, because it can achieve throughput, rate and delay guarantees. In this article, we propose a practical scheduling scheme called the weighted quantitative queue longest and least scheduled first, shorted by WQ2L2SF. The main features of this scheduling algorithm are in its easier implementation in hardware, as the time complexities is only O(logN), and fewer hardware resources are needed than existing schemes. We proved that under a speedup factor of 2, our algorithm can achieve 100% throughput. Finally, Simulation results show that our scheduling scheme can achieve good delay and stability performances