Pub Date : 2001-05-29DOI: 10.1109/HPSR.2001.923610
K. Yamakoshi, K. Nakai, N. Matsuura, E. Oki, N. Yamanaka
This paper describes the hardware architecture and performance of a 5-Tb/s frame-based (FB) ATM switching system OPTIMA-P (OPTically Interconnected Distributed Multi-stage Tb/s-ATM switching Network Architecture -2). First, we describe the switch architecture, which is a non-blocking 3-stage switch using optical wavelength division multiplexing (WDM) and dynamic bandwidth sharing technique. Variable-length FB-ATM cells, which can encapsulate IP packets, can be switched while keeping throughput fairness. Second, we propose a scheduling technique for variable-length packets used in the WDM system which selects packets from the eight wavelengths. We show that the scheduling technique can keep the fairness and has a small delay for variable-length packets. A pipelined-scheduling technique for variable-length packets is also proposed for high-speed and large-capacity switching system.
{"title":"5-Tb/s frame-based ATM switching system using 2.5-Gb/s/spl times/8-/spl lambda/ optical switching technique","authors":"K. Yamakoshi, K. Nakai, N. Matsuura, E. Oki, N. Yamanaka","doi":"10.1109/HPSR.2001.923610","DOIUrl":"https://doi.org/10.1109/HPSR.2001.923610","url":null,"abstract":"This paper describes the hardware architecture and performance of a 5-Tb/s frame-based (FB) ATM switching system OPTIMA-P (OPTically Interconnected Distributed Multi-stage Tb/s-ATM switching Network Architecture -2). First, we describe the switch architecture, which is a non-blocking 3-stage switch using optical wavelength division multiplexing (WDM) and dynamic bandwidth sharing technique. Variable-length FB-ATM cells, which can encapsulate IP packets, can be switched while keeping throughput fairness. Second, we propose a scheduling technique for variable-length packets used in the WDM system which selects packets from the eight wavelengths. We show that the scheduling technique can keep the fairness and has a small delay for variable-length packets. A pipelined-scheduling technique for variable-length packets is also proposed for high-speed and large-capacity switching system.","PeriodicalId":308964,"journal":{"name":"2001 IEEE Workshop on High Performance Switching and Routing (IEEE Cat. No.01TH8552)","volume":"284 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122989234","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2001-05-29DOI: 10.1109/HPSR.2001.923652
Xiaohong Jiang, M.M. Khandker, S. Horiguchi
Vertical stacking is a novel scheme in creating nonblocking multistage interconnection networks (MINs). Available results indicate that under the crosstalk-free constraint, the hardware cost is high for a strictly nonblocking optical MIN based on the scheme. Packing strategy is a commonly used routing strategy for reducing network hardware. In this paper, we look into the nonblocking conditions for crosstalk-free optical MINs under packing strategy. It is shown that packing strategy is very efficient in reducing the hardware cost of a nonblocking optical MIN under the crosstalk-free constraint. The results in the paper will be useful in designing optical MINs.
{"title":"Nonblocking optical MINs under crosstalk-free constraint","authors":"Xiaohong Jiang, M.M. Khandker, S. Horiguchi","doi":"10.1109/HPSR.2001.923652","DOIUrl":"https://doi.org/10.1109/HPSR.2001.923652","url":null,"abstract":"Vertical stacking is a novel scheme in creating nonblocking multistage interconnection networks (MINs). Available results indicate that under the crosstalk-free constraint, the hardware cost is high for a strictly nonblocking optical MIN based on the scheme. Packing strategy is a commonly used routing strategy for reducing network hardware. In this paper, we look into the nonblocking conditions for crosstalk-free optical MINs under packing strategy. It is shown that packing strategy is very efficient in reducing the hardware cost of a nonblocking optical MIN under the crosstalk-free constraint. The results in the paper will be useful in designing optical MINs.","PeriodicalId":308964,"journal":{"name":"2001 IEEE Workshop on High Performance Switching and Routing (IEEE Cat. No.01TH8552)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129691735","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2001-05-29DOI: 10.1109/HPSR.2001.923598
P. Gadiraju, H. Mouftah
With the increased transmission capacity in WDM networks a single failure in the network (like a fiber cut or equipment failure) may result in enormous data loss. Therefore, resilience from a component or optical link failure is essential to provide uninterrupted communication between any two nodes in the network. Several techniques exist to provide reliable services in the occurrence of failures. The most common protection technique involves assigning spare capacity. This study focuses on providing channel protection against single failures for networks with or without wavelength converters. In this study a new protection technique is introduced based on the 1:N spare capacity assignment and has been incorporated in RWANCO (routing and wavelength assignment and network cost optimization). The method involves changing the protection wavelength after each failure. Also, the study investigates the effect of this changing wavelength which introduces some blocking in the network. We also evaluate the protection switching time for the different cases.
{"title":"Channel protection in WDM mesh networks","authors":"P. Gadiraju, H. Mouftah","doi":"10.1109/HPSR.2001.923598","DOIUrl":"https://doi.org/10.1109/HPSR.2001.923598","url":null,"abstract":"With the increased transmission capacity in WDM networks a single failure in the network (like a fiber cut or equipment failure) may result in enormous data loss. Therefore, resilience from a component or optical link failure is essential to provide uninterrupted communication between any two nodes in the network. Several techniques exist to provide reliable services in the occurrence of failures. The most common protection technique involves assigning spare capacity. This study focuses on providing channel protection against single failures for networks with or without wavelength converters. In this study a new protection technique is introduced based on the 1:N spare capacity assignment and has been incorporated in RWANCO (routing and wavelength assignment and network cost optimization). The method involves changing the protection wavelength after each failure. Also, the study investigates the effect of this changing wavelength which introduces some blocking in the network. We also evaluate the protection switching time for the different cases.","PeriodicalId":308964,"journal":{"name":"2001 IEEE Workshop on High Performance Switching and Routing (IEEE Cat. No.01TH8552)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128574972","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2001-05-29DOI: 10.1109/HPSR.2001.923670
Ying Jiang, M. Hamdi
Virtual output queuing (VOQ) is a practical and high-performance packet switch architecture. There are many simple iterative arbitration algorithms proposed for the VOQ architecture. We investigate in this paper the performance of various such algorithms and based on the analysis of pointer desynchronization effect, we propose a group of new arbitration algorithms, called SRR (static round robin matching) which perform pretty well under various traffic models and are easy to implement in hardware.
{"title":"A fully desynchronized round-robin matching scheduler for a VOQ packet switch architecture","authors":"Ying Jiang, M. Hamdi","doi":"10.1109/HPSR.2001.923670","DOIUrl":"https://doi.org/10.1109/HPSR.2001.923670","url":null,"abstract":"Virtual output queuing (VOQ) is a practical and high-performance packet switch architecture. There are many simple iterative arbitration algorithms proposed for the VOQ architecture. We investigate in this paper the performance of various such algorithms and based on the analysis of pointer desynchronization effect, we propose a group of new arbitration algorithms, called SRR (static round robin matching) which perform pretty well under various traffic models and are easy to implement in hardware.","PeriodicalId":308964,"journal":{"name":"2001 IEEE Workshop on High Performance Switching and Routing (IEEE Cat. No.01TH8552)","volume":"169 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128631744","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2001-05-29DOI: 10.1109/HPSR.2001.923606
S. Zeng, N. Uzun, S. Papavassiliou
This paper presents an improved fair traffic shaping algorithm for high speed packet switches. This algorithm introduces a resorting process to resolve the contention of multiple cells which are all eligible for transmission in the traffic shaper. By using the resorting process based on the each connection's rate, we can give better fairness and flexibility in the bandwidth assignment for connections with wide range of rates. Compared with the departure event driven traffic shaping algorithm (DEDTS), we can reduce the shaping delay for high rate connections without increasing the total available bandwidth. From simulation experiments, we show that the proposed algorithm can implement the natural prioritization for connections based on their individual rates, and improve the fairness in the bandwidth assignment. A hardware implementation architecture for this algorithm is also given.
{"title":"An improved fair traffic shaping algorithm for high speed packet switches","authors":"S. Zeng, N. Uzun, S. Papavassiliou","doi":"10.1109/HPSR.2001.923606","DOIUrl":"https://doi.org/10.1109/HPSR.2001.923606","url":null,"abstract":"This paper presents an improved fair traffic shaping algorithm for high speed packet switches. This algorithm introduces a resorting process to resolve the contention of multiple cells which are all eligible for transmission in the traffic shaper. By using the resorting process based on the each connection's rate, we can give better fairness and flexibility in the bandwidth assignment for connections with wide range of rates. Compared with the departure event driven traffic shaping algorithm (DEDTS), we can reduce the shaping delay for high rate connections without increasing the total available bandwidth. From simulation experiments, we show that the proposed algorithm can implement the natural prioritization for connections based on their individual rates, and improve the fairness in the bandwidth assignment. A hardware implementation architecture for this algorithm is also given.","PeriodicalId":308964,"journal":{"name":"2001 IEEE Workshop on High Performance Switching and Routing (IEEE Cat. No.01TH8552)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130490575","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2001-05-29DOI: 10.1109/HPSR.2001.923617
A. Faragó
We investigate the problem of how to combine arbitrary link blocking models into a network level model, without assuming the probabilistic independence of links, as opposed to previous approaches. The significance of the problem stems from a number of networking scenarios where the classical link independence assumption fails. We show that with a new system of fixed point equations it is possible to exactly compute the link and route availability probabilities and the exact network revenue, even in the presence of strongly dependent links. As a by-product of the method we also obtain a generalization of the product, and summation bounds of Whitt (1985) for arbitrary multiservice traffic scenarios.
{"title":"A general method for the blocking analysis of networks with dependent links","authors":"A. Faragó","doi":"10.1109/HPSR.2001.923617","DOIUrl":"https://doi.org/10.1109/HPSR.2001.923617","url":null,"abstract":"We investigate the problem of how to combine arbitrary link blocking models into a network level model, without assuming the probabilistic independence of links, as opposed to previous approaches. The significance of the problem stems from a number of networking scenarios where the classical link independence assumption fails. We show that with a new system of fixed point equations it is possible to exactly compute the link and route availability probabilities and the exact network revenue, even in the presence of strongly dependent links. As a by-product of the method we also obtain a generalization of the product, and summation bounds of Whitt (1985) for arbitrary multiservice traffic scenarios.","PeriodicalId":308964,"journal":{"name":"2001 IEEE Workshop on High Performance Switching and Routing (IEEE Cat. No.01TH8552)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129212549","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2001-05-29DOI: 10.1109/HPSR.2001.923651
T. Michalareas, Lionel Sacks
In this paper we examine the steady state behaviour of the link state routing protocol and the AntNet algorithm in simple topologies under uniform load. We show that under our scenario conditions both algorithms behave in the same operational region. We discuss also the applied statistical methodology and the necessary measurements for throughput behaviour in steady state.
{"title":"Link-state and ant-like algorithm behaviour for single-constrained routing","authors":"T. Michalareas, Lionel Sacks","doi":"10.1109/HPSR.2001.923651","DOIUrl":"https://doi.org/10.1109/HPSR.2001.923651","url":null,"abstract":"In this paper we examine the steady state behaviour of the link state routing protocol and the AntNet algorithm in simple topologies under uniform load. We show that under our scenario conditions both algorithms behave in the same operational region. We discuss also the applied statistical methodology and the necessary measurements for throughput behaviour in steady state.","PeriodicalId":308964,"journal":{"name":"2001 IEEE Workshop on High Performance Switching and Routing (IEEE Cat. No.01TH8552)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126376819","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2001-05-29DOI: 10.1109/HPSR.2001.923641
Wei Wu, Yong Ren, X. Shan
Most applications can adapt their coding techniques and sending rates according to the network congestion and the resource needed can be provided at the beginning of the transmission. So the traditional DiffServ model is too rigid to them. In this paper, we are seeking a balance between the relative DiffServ and the absolute DiffServ and propose a new Diffserv model, a relative differentiated service model with admission control, which suits the adaptive application. By providing the proportional differentiated services in core routers and loss-rate based CAC control in edge routers, we can make both the network and the users adaptive: the network is adaptive to the traffic load and the users is adaptive to the network congestion. This model is promising to the elastic but unpredictable traffic, such as IP telephony or other multimedia applications.
{"title":"Forwarding the balance between absolute and relative: a new differentiated services model for adaptive traffic","authors":"Wei Wu, Yong Ren, X. Shan","doi":"10.1109/HPSR.2001.923641","DOIUrl":"https://doi.org/10.1109/HPSR.2001.923641","url":null,"abstract":"Most applications can adapt their coding techniques and sending rates according to the network congestion and the resource needed can be provided at the beginning of the transmission. So the traditional DiffServ model is too rigid to them. In this paper, we are seeking a balance between the relative DiffServ and the absolute DiffServ and propose a new Diffserv model, a relative differentiated service model with admission control, which suits the adaptive application. By providing the proportional differentiated services in core routers and loss-rate based CAC control in edge routers, we can make both the network and the users adaptive: the network is adaptive to the traffic load and the users is adaptive to the network congestion. This model is promising to the elastic but unpredictable traffic, such as IP telephony or other multimedia applications.","PeriodicalId":308964,"journal":{"name":"2001 IEEE Workshop on High Performance Switching and Routing (IEEE Cat. No.01TH8552)","volume":"484 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115878268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2001-05-29DOI: 10.1109/HPSR.2001.923660
Jun Wang, Klara Nahrstedt
The invention and evolution of the dense wavelength division multiplexing (DWDM) technology has brought a breakthrough to high-speed networks, and it has put a lot of pressure on research in the area of IP router to catch up. Besides, with up-coming quality of service (QoS) requirements raised by a wide range of communication-intensive, real-time multimedia applications, the next-generation IP routers should be QoS-capable. Limited by Moore's law, one possible solution is to introduce parallelism as well as the differentiated service (DiffServ) scheme [5, 11] into the router architecture to provide QoS provision at a high speed and a low cost. We propose a novel architecture called the high-performance QoS-capable IP router (HPQR). We address one key design issue in our architecture-the distribution of IP packets to multiple independent routing agents so that the workload at routing agents is balanced and the packet ordering is preserved. We introduce the enhanced hash-based distributing scheme (EHDS) as the solution. Simulations are carried out to study the effectiveness of EHDS. The results show that EHDS does meet our design goals very well.
{"title":"Parallel IP packet forwarding for tomorrow's IP routers","authors":"Jun Wang, Klara Nahrstedt","doi":"10.1109/HPSR.2001.923660","DOIUrl":"https://doi.org/10.1109/HPSR.2001.923660","url":null,"abstract":"The invention and evolution of the dense wavelength division multiplexing (DWDM) technology has brought a breakthrough to high-speed networks, and it has put a lot of pressure on research in the area of IP router to catch up. Besides, with up-coming quality of service (QoS) requirements raised by a wide range of communication-intensive, real-time multimedia applications, the next-generation IP routers should be QoS-capable. Limited by Moore's law, one possible solution is to introduce parallelism as well as the differentiated service (DiffServ) scheme [5, 11] into the router architecture to provide QoS provision at a high speed and a low cost. We propose a novel architecture called the high-performance QoS-capable IP router (HPQR). We address one key design issue in our architecture-the distribution of IP packets to multiple independent routing agents so that the workload at routing agents is balanced and the packet ordering is preserved. We introduce the enhanced hash-based distributing scheme (EHDS) as the solution. Simulations are carried out to study the effectiveness of EHDS. The results show that EHDS does meet our design goals very well.","PeriodicalId":308964,"journal":{"name":"2001 IEEE Workshop on High Performance Switching and Routing (IEEE Cat. No.01TH8552)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132075507","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2001-05-29DOI: 10.1109/HPSR.2001.923643
B. Gaidioz, P. Primet, B. Tourancheau
Real-time traffic in the Internet is growing and the best-effort model has not been designed for such challenging streams. The DiffServ (DS) architecture, which permits one to aggregate traffic flows into classes with a common forwarding requirement, is a quality of service (QoS) solution for IP networks. But most of the DS models provide classes getting better performance than others, calling for pricing differentiation and possibly admission control. We explore the possibility of providing a simple, robust and pricing-free QoS solution by practising "differentiated fairness": different classes have equivalent performance according to their specific needs. No admission control is required, no absolute guarantee is provided. This "soft" model has been implemented and tested. Experiments shows that the scheduler can balance resource sharing so that "differentiated fairness" is obtained.
{"title":"Differentiated fairness: service model and implementation","authors":"B. Gaidioz, P. Primet, B. Tourancheau","doi":"10.1109/HPSR.2001.923643","DOIUrl":"https://doi.org/10.1109/HPSR.2001.923643","url":null,"abstract":"Real-time traffic in the Internet is growing and the best-effort model has not been designed for such challenging streams. The DiffServ (DS) architecture, which permits one to aggregate traffic flows into classes with a common forwarding requirement, is a quality of service (QoS) solution for IP networks. But most of the DS models provide classes getting better performance than others, calling for pricing differentiation and possibly admission control. We explore the possibility of providing a simple, robust and pricing-free QoS solution by practising \"differentiated fairness\": different classes have equivalent performance according to their specific needs. No admission control is required, no absolute guarantee is provided. This \"soft\" model has been implemented and tested. Experiments shows that the scheduler can balance resource sharing so that \"differentiated fairness\" is obtained.","PeriodicalId":308964,"journal":{"name":"2001 IEEE Workshop on High Performance Switching and Routing (IEEE Cat. No.01TH8552)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133057310","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：