Journal of the Korea Institute of Information and Communication Engineering (한국정보통신학회논문지)

The Korea Institute of Information and Commucation Engineering (한국정보통신학회)
권호 표지

Non-Work Conserving Round Robin Schedulers

비 작업보존형 라운드로빈 스케줄러

  • 정진우 (상명대학교 소프트웨어학부)
  • Published : 2005.12.01
Download PDF
⟨ Previous Next ⟩

Abstract

There have been numerous researches regarding the QoS guarantee in packet switching networks. IntServs, based on a signaling mechanism and scheduling algorithms, suggesting promising solutions, yet has the crucial complexity problem so that not enough real implementations has been witnessed. Flow aggregation is suggested recently to overcome this issue. In order to aggregated flows fairly so that the latency of the aggregated flows is bound, however, a non-work conserving scheduler is necessary, which is not very popular because of its another inherent complexity. We suggest a non-work conserving scheduler, the Round Robin with Virtual Flow (RRVF), which is a variation of the popular Deficit Round Robin (DRR). We study the latency of the RRVF, and observe that the non-work conserving nature of the RRVF yields a slight disadvantage in terms of the latency, but after the aggregation the latency is greatly reduced, so that e combined latency is reduced. We conclude that the flow aggregation through RRVF can actually reduce the complexity of the bandwidth allocation as well as the overall latency within a network.

패킷 스위칭 네트워크에서의 QoS 보장을 위한 많은 연구가 지난 10여 년간 진행되었다. 이들 중 많은 수가 IntServs 기반의 플로우별 대역폭 할당과 보장을 위한 여러 가지 시그널링, 스케줄링 방법에 관한 것이나, 구현의 복잡성으로 인해 실제 네트워크에서 실현된 예가 드물다. 이러한 복잡성을 극복하기 위해서 최근 플로우 통합 (Flow Aggregation)이 제안된 바 있다. 통합된 플로우 기반의 스케줄링을 통해 지연시간을 보장해주기 위해서는 플로우 간의 공정한 통합이 요구되며, 이를 위해서 스케줄러가 비 작업보존방식으로 동작해야 한다. 본고에서는 가장 간단하면서 별리 쓰이는 스케줄링 기법인 Deficit Round Robin을 변형한, 비 작업보존 방식의 Round Robin with Virtual Flow (RRVF)를 제안하고 이를 바탕으로 공정한 플로우간 통합을 시도하였다. RRVF가 보장하는 지연시간 최대치를 연구하였으며 이를 적용한 플로우 통합시의 지연시간 최대치를 구하였다. 이를 통해 RRVF로 플로우를 통합하는 경우 대역폭 할당이 간단해짐과 동시에 네트워크 전체의 지연시간도 줄어드는 것을 알아내었다.

Keywords

References

  1. A. Demers, S. Keshav, and S. Shenker, 'Analysis and Simulation of a Fair Queuing Algorithm,' in ACM SIGCOMM, pp. 1-12, Sep. 1989
  2. A. K. Parekh and R. G. Gallager, 'A generalized processor sharing approach to flow control in integrated services networks: The singlenode case,' IEEE/ACM Trans. Networking, vol. 1, no. 3, pp. 344 - 357, June 1993 https://doi.org/10.1109/90.234856
  3. S. Golestani, 'A Self-clocked Fair Queueing Scheme for Broadband Applications', in IEEE INFOCOM'94, 1994
  4. M. Shreedhar and G. Varghese, 'Efficient fair queueing using deficit round-robin,' IEEE/ACM Trans. Networking, vol. 4, no. 3, pp. 375 - 385, June 1996 https://doi.org/10.1109/90.502236
  5. L. Lenzini, E. Mingozzi, and G. Shea, 'Tradeoffs between low complexity, low latency, and fairness with deficit round-robin schedulers,' lEEE/ACM Trans. Networking, vol. 12, no. 4, pp. 681 - 693, Aug. 2004 https://doi.org/10.1109/TNET.2004.833131
  6. J. A. Cobb, 'Preserving Quality of Service guarantees in spite of flow aggregation', IEEE/ACM Trans. Networking, vol. 10, no. 1, Feb. 2002
  7. S. S. Kanhere and H. Sethu, 'On the latency bound of deficit round robin', in Proceedings of the ICCCN, Miami, Oct. 2002
  8. D. Stiliadis and A. Varma, 'Latency-Rate servers: A general model for analysis of traffic scheduling algorithms', IEEE/ACM Trans. Networking, vol. 6, no. 5 Oct. 1998
  9. I. Stoica and H. Zhang, 'Providing guaranteed services without per flow management,' in Proc. ACM SIGCOMM, pp. 81 - 94, Aug. 1999