Pub Date : 2008-05-15DOI: 10.1109/HSPR.2008.4734429
Feng Wang, M. Hamdi
With the constantly increasing Internet traffic, buffers are becoming major bottlenecks for todaypsilas high-end routers. In particular, router buffers are required to have both high speed and large capacities, which are hard to build with current single memory technology, such as SRAM or DRAM. A general approach is to make a combination of the SRAM and DRAM and exploit advantages from both. The main obstacle is to find a way matching the speed gap between them. And the requirement to maintain multiple flows in the system further complicates the problem. In this paper, we first investigate previous solutions that use different access granularities to match the speed gap. We point out their intrinsic scaling problems when the number of flows increases. Then, we propose to use parallelism to match the speed gap. Numerical studies and simulations both show that our proposal can theoretically support any number of flows in the router with just little SRAM under practical traffic. In addition, the memory management algorithm is also more scalable compared to those in previous solutions.
{"title":"Matching the speed gap between SRAM and DRAM","authors":"Feng Wang, M. Hamdi","doi":"10.1109/HSPR.2008.4734429","DOIUrl":"https://doi.org/10.1109/HSPR.2008.4734429","url":null,"abstract":"With the constantly increasing Internet traffic, buffers are becoming major bottlenecks for todaypsilas high-end routers. In particular, router buffers are required to have both high speed and large capacities, which are hard to build with current single memory technology, such as SRAM or DRAM. A general approach is to make a combination of the SRAM and DRAM and exploit advantages from both. The main obstacle is to find a way matching the speed gap between them. And the requirement to maintain multiple flows in the system further complicates the problem. In this paper, we first investigate previous solutions that use different access granularities to match the speed gap. We point out their intrinsic scaling problems when the number of flows increases. Then, we propose to use parallelism to match the speed gap. Numerical studies and simulations both show that our proposal can theoretically support any number of flows in the router with just little SRAM under practical traffic. In addition, the memory management algorithm is also more scalable compared to those in previous solutions.","PeriodicalId":130484,"journal":{"name":"2008 International Conference on High Performance Switching and Routing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129962740","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 : 2008-05-15DOI: 10.1109/HSPR.2008.4734418
T. Orawiwattanakul, Yusheng Ji, Yongbing Zhang, Jie Li
Fairness in OBS networks is becoming a significant issue due to the network high loss characteristic. We propose a rate fairness preemption (RFP) scheme that can provide rate fairness in OBS networks. The RFP scheme adaptively allocates a fair amount of bandwidth to all connections and uses the preemption technique to provide this fairness. We present the analytical burst loss model for RFP implementation. Analysis and simulation results show that the RFP scheme effectively provides rate fairness in OBS networks.
{"title":"Preemption scheme for providing Rate Fairness in optical burst switching networks","authors":"T. Orawiwattanakul, Yusheng Ji, Yongbing Zhang, Jie Li","doi":"10.1109/HSPR.2008.4734418","DOIUrl":"https://doi.org/10.1109/HSPR.2008.4734418","url":null,"abstract":"Fairness in OBS networks is becoming a significant issue due to the network high loss characteristic. We propose a rate fairness preemption (RFP) scheme that can provide rate fairness in OBS networks. The RFP scheme adaptively allocates a fair amount of bandwidth to all connections and uses the preemption technique to provide this fairness. We present the analytical burst loss model for RFP implementation. Analysis and simulation results show that the RFP scheme effectively provides rate fairness in OBS networks.","PeriodicalId":130484,"journal":{"name":"2008 International Conference on High Performance Switching and Routing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129987451","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 : 2008-05-15DOI: 10.1109/HSPR.2008.4734427
M. Chiang, Zabih Ghassemlooy, W. P. Ng, H. L. Minh, C. Lu
The paper presents a new node architecture for an all-optical packet router employing multiple pulse position modulation (PPM) routing table with a hybrid packet header correlation scheme. Most existing routing tables within a node contain a large number of entries, thus resulting in a long packet header address correlation time before delivering the incoming packet to its destination. In the proposed multiple PPM routing tables (PPRTs) the packet header address is based on the binary and PPM formats which leads to a much reduced routing table size. The packet header address correlation is carried out using only a single optical AND gate, thus offering reduced system complexity. It is also shown that the proposed scheme offers unicast/multi-cast/broadcast transmitting capabilities. The propose scheme is simulated and its characteristics are investigated. The output inter-channel crosstalk (CXT) of up to -18 dB and output packet power fluctuation of 2 dB have been achieved, which largely depend on the guard time between the arriving packets.
{"title":"Ultra-fast all-optical packet-switched routing with a hybrid header address correlation Scheme","authors":"M. Chiang, Zabih Ghassemlooy, W. P. Ng, H. L. Minh, C. Lu","doi":"10.1109/HSPR.2008.4734427","DOIUrl":"https://doi.org/10.1109/HSPR.2008.4734427","url":null,"abstract":"The paper presents a new node architecture for an all-optical packet router employing multiple pulse position modulation (PPM) routing table with a hybrid packet header correlation scheme. Most existing routing tables within a node contain a large number of entries, thus resulting in a long packet header address correlation time before delivering the incoming packet to its destination. In the proposed multiple PPM routing tables (PPRTs) the packet header address is based on the binary and PPM formats which leads to a much reduced routing table size. The packet header address correlation is carried out using only a single optical AND gate, thus offering reduced system complexity. It is also shown that the proposed scheme offers unicast/multi-cast/broadcast transmitting capabilities. The propose scheme is simulated and its characteristics are investigated. The output inter-channel crosstalk (CXT) of up to -18 dB and output packet power fluctuation of 2 dB have been achieved, which largely depend on the guard time between the arriving packets.","PeriodicalId":130484,"journal":{"name":"2008 International Conference on High Performance Switching and Routing","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124470611","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 : 2008-05-15DOI: 10.1109/HSPR.2008.4734428
Xiangjie Ma, Julong Lan, Baisheng Zhang
Clos packet switching networks are the next step in scaling current Crossbar switches to large number of ports. The novel central-stage buffered Clos-network (CBC network) has the merit of fully memory-sharing among all inputs and outputs and was proved to emulate an FCFS-OQ switch with the central module count (i.e. m) about four times of the input modules count (i.e. n). This leads to high hardware complexity when designing a CBC network with high radices. This paper studies the graphic model of the CBC network, proposes analytical methods including the stable path set and the stable vertex pair set, and demonstrates that the CBC network can match an FCFS-OQ switch with the central module count m>=n. By comparison, the number of the central modules in our result is only a quarter of that in previous results, and therefore the implementation of a CBC network become much simpler in practical routers.
{"title":"Emulating output queueing with the Central-stage Buffered Clos packet switching network","authors":"Xiangjie Ma, Julong Lan, Baisheng Zhang","doi":"10.1109/HSPR.2008.4734428","DOIUrl":"https://doi.org/10.1109/HSPR.2008.4734428","url":null,"abstract":"Clos packet switching networks are the next step in scaling current Crossbar switches to large number of ports. The novel central-stage buffered Clos-network (CBC network) has the merit of fully memory-sharing among all inputs and outputs and was proved to emulate an FCFS-OQ switch with the central module count (i.e. m) about four times of the input modules count (i.e. n). This leads to high hardware complexity when designing a CBC network with high radices. This paper studies the graphic model of the CBC network, proposes analytical methods including the stable path set and the stable vertex pair set, and demonstrates that the CBC network can match an FCFS-OQ switch with the central module count m>=n. By comparison, the number of the central modules in our result is only a quarter of that in previous results, and therefore the implementation of a CBC network become much simpler in practical routers.","PeriodicalId":130484,"journal":{"name":"2008 International Conference on High Performance Switching and Routing","volume":"95 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127100907","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 : 2008-05-15DOI: 10.1109/HSPR.2008.4734423
I. Takouna, R. Rojas-Cessa
Routing schemes combined with link-state detection mechanisms can be used to recover connection paths or network connectivity under the cases of link and node failures. Some schemes have large computation and state-update overhead as re-routing is performed throughout the network. Moreover, as soon as new routes are found, some link may become congested with flows under recovery. This congestion may add extra recovery delays and even further link or node failures. In this paper, we propose proactive routing recovery schemes that perform rerouting on links affected by the failure, therefore, minimizing the computation overhead. Congestion avoidance is also achieved in these schemes by calculating the distribution of re-routed traffic in a proactive fashion. We compare our proposed schemes with the open shortest path first (OSPF) scheme and show that our schemes can provide higher utilization of links and nodes for large networks in post-recovery. We show simulation results under link and node failures.
{"title":"Routing schemes for network recovery under link and node failures","authors":"I. Takouna, R. Rojas-Cessa","doi":"10.1109/HSPR.2008.4734423","DOIUrl":"https://doi.org/10.1109/HSPR.2008.4734423","url":null,"abstract":"Routing schemes combined with link-state detection mechanisms can be used to recover connection paths or network connectivity under the cases of link and node failures. Some schemes have large computation and state-update overhead as re-routing is performed throughout the network. Moreover, as soon as new routes are found, some link may become congested with flows under recovery. This congestion may add extra recovery delays and even further link or node failures. In this paper, we propose proactive routing recovery schemes that perform rerouting on links affected by the failure, therefore, minimizing the computation overhead. Congestion avoidance is also achieved in these schemes by calculating the distribution of re-routed traffic in a proactive fashion. We compare our proposed schemes with the open shortest path first (OSPF) scheme and show that our schemes can provide higher utilization of links and nodes for large networks in post-recovery. We show simulation results under link and node failures.","PeriodicalId":130484,"journal":{"name":"2008 International Conference on High Performance Switching and Routing","volume":"878 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131521211","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 : 2008-05-15DOI: 10.1109/HSPR.2008.4734452
A. Bianco, J. Finochietto, Marco Modesti, F. Neri
iSCSI is a protocol standardized by IETF that runs on a TCP/IP network to transfer standard SCSI commands. In this paper, we evaluate the performance of iSCSI using standard PCs running a software implementation of the protocol, with the aim of assessing the performance of low-cost distributed storage solutions. First, we compare the performance when using a local disk with those of a virtualized disk connected through a standard gigabit Ethernet LAN during file I/O operations. Then, we emulate the characteristics of a WAN/MAN setup by using a software emulator within the Linux kernel to generate delays and packet losses, and evaluate the ability of iSCSI to offer disaster recovery solutions over high-speed long-distance links. We found that, even if TCP can be tuned to support links with large bandwidth-delay products, iSCSI does not reach good performance due to its windowing scheme that is not large enough, in the used software implementation, to support long-distance links. Finally, we test the iSCSI behavior when other types of traffic share network links with the iSCSI protocol.
{"title":"Distributed storage on networks of Linux PCs using the iSCSI protocol","authors":"A. Bianco, J. Finochietto, Marco Modesti, F. Neri","doi":"10.1109/HSPR.2008.4734452","DOIUrl":"https://doi.org/10.1109/HSPR.2008.4734452","url":null,"abstract":"iSCSI is a protocol standardized by IETF that runs on a TCP/IP network to transfer standard SCSI commands. In this paper, we evaluate the performance of iSCSI using standard PCs running a software implementation of the protocol, with the aim of assessing the performance of low-cost distributed storage solutions. First, we compare the performance when using a local disk with those of a virtualized disk connected through a standard gigabit Ethernet LAN during file I/O operations. Then, we emulate the characteristics of a WAN/MAN setup by using a software emulator within the Linux kernel to generate delays and packet losses, and evaluate the ability of iSCSI to offer disaster recovery solutions over high-speed long-distance links. We found that, even if TCP can be tuned to support links with large bandwidth-delay products, iSCSI does not reach good performance due to its windowing scheme that is not large enough, in the used software implementation, to support long-distance links. Finally, we test the iSCSI behavior when other types of traffic share network links with the iSCSI protocol.","PeriodicalId":130484,"journal":{"name":"2008 International Conference on High Performance Switching and Routing","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129512767","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 : 2008-05-15DOI: 10.1109/HSPR.2008.4734434
Longjiang Li, Y. Mao, Yonggang Li
IETF mobile IP version 6 and its fast handover protocol provide transport layer connection survivability when a node moves from one access router to another by selecting a new care-of address (NCoA) instead of changing the home address. The aDAD(advance duplicate address detection) protocol can almost eliminate the latency needed for the address configuration and confirmation from the whole IPv6 handover latency, but uses excessively a lot of network resources, such as bandwidth and IO load, to generate NCoAs and check their uniqueness. This paper proposes agent-based DAD, a novel approach for efficiently provisioning NCoAs. The proposed scheme can reduce the required network resources significantly for generating these NCoAs, and keeping their uniqueness but does not change the basic behavior of aDAD.
{"title":"Resource-efficient care-of address provisioning for seamless IPv6 mobility support","authors":"Longjiang Li, Y. Mao, Yonggang Li","doi":"10.1109/HSPR.2008.4734434","DOIUrl":"https://doi.org/10.1109/HSPR.2008.4734434","url":null,"abstract":"IETF mobile IP version 6 and its fast handover protocol provide transport layer connection survivability when a node moves from one access router to another by selecting a new care-of address (NCoA) instead of changing the home address. The aDAD(advance duplicate address detection) protocol can almost eliminate the latency needed for the address configuration and confirmation from the whole IPv6 handover latency, but uses excessively a lot of network resources, such as bandwidth and IO load, to generate NCoAs and check their uniqueness. This paper proposes agent-based DAD, a novel approach for efficiently provisioning NCoAs. The proposed scheme can reduce the required network resources significantly for generating these NCoAs, and keeping their uniqueness but does not change the basic behavior of aDAD.","PeriodicalId":130484,"journal":{"name":"2008 International Conference on High Performance Switching and Routing","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131336477","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 : 2008-05-15DOI: 10.1109/HSPR.2008.4734445
Zhi-hu Tan, C. Xie, Ji-guang Wan
RAID has been a key component for constructing enterprise storage system. The target module supports host for the access interface of storage devices. This paper discusses the function of the target module based on a prototype of fiber channel RAID system. Further analyses the internal architecture of target module and gives an elaboration on the various workflows of the read/write access. Finally, we use some general benchmarks to evaluate the performance of this fiber channel target module and RAID system. Testing results reveals a moderately successful performance of the fiber channel target module.
{"title":"Performance benchmark for target module in Fiber Channel RAID systems","authors":"Zhi-hu Tan, C. Xie, Ji-guang Wan","doi":"10.1109/HSPR.2008.4734445","DOIUrl":"https://doi.org/10.1109/HSPR.2008.4734445","url":null,"abstract":"RAID has been a key component for constructing enterprise storage system. The target module supports host for the access interface of storage devices. This paper discusses the function of the target module based on a prototype of fiber channel RAID system. Further analyses the internal architecture of target module and gives an elaboration on the various workflows of the read/write access. Finally, we use some general benchmarks to evaluate the performance of this fiber channel target module and RAID system. Testing results reveals a moderately successful performance of the fiber channel target module.","PeriodicalId":130484,"journal":{"name":"2008 International Conference on High Performance Switching and Routing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129258991","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 : 2008-05-15DOI: 10.1109/HSPR.2008.4734450
S. Hauger, M. Scharf, J. Kogel, C. Suriyajan
The quick-start extension of the transmission control protocol (TCP), as well as the explicit control protocol (XCP), are experimental congestion control schemes that use router feedback to overcome limitations of TCPpsilas standard mechanisms. Both approaches require additional packet processing in every router and therefore raise the question whether, and how, this can be achieved in high-speed routers. This paper studies the realization complexity of the quick-start and XCP router functions on a network processor. We show that in both cases synchronization issues among parallel processing entities have to be considered, and that this affects the router performance. We develop and compare different synchronization mechanisms for highly parallel packet processing. Our prototype implementation on an Intel IXP network processor allows to quantify the impact on throughput and delay caused by the additional packet processing in the fast path. The measurements reveal that quick-start and XCP processing is feasible at multiple Gbit/s line speed, with quick-start being simpler to scale.
{"title":"Quick-Start and XCP on a network processor: Implementation issues and performance evaluation","authors":"S. Hauger, M. Scharf, J. Kogel, C. Suriyajan","doi":"10.1109/HSPR.2008.4734450","DOIUrl":"https://doi.org/10.1109/HSPR.2008.4734450","url":null,"abstract":"The quick-start extension of the transmission control protocol (TCP), as well as the explicit control protocol (XCP), are experimental congestion control schemes that use router feedback to overcome limitations of TCPpsilas standard mechanisms. Both approaches require additional packet processing in every router and therefore raise the question whether, and how, this can be achieved in high-speed routers. This paper studies the realization complexity of the quick-start and XCP router functions on a network processor. We show that in both cases synchronization issues among parallel processing entities have to be considered, and that this affects the router performance. We develop and compare different synchronization mechanisms for highly parallel packet processing. Our prototype implementation on an Intel IXP network processor allows to quantify the impact on throughput and delay caused by the additional packet processing in the fast path. The measurements reveal that quick-start and XCP processing is feasible at multiple Gbit/s line speed, with quick-start being simpler to scale.","PeriodicalId":130484,"journal":{"name":"2008 International Conference on High Performance Switching and Routing","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114877514","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 : 2008-05-15DOI: 10.1109/HSPR.2008.4734439
J. Lopez, J. Pedersen, R. C. Rumín, O. Madsen
The topological routing over N2R structures has previously been studied and implemented using different techniques. A first approach was achieved obtaining the best trade off between path length vs. path completion time for the shortest path between any pair of nodes. This paper introduces protection against failures by modifying the previous algorithm implementing the option of routing a packet using a second independent path. The goal is to prove that there is an easy and efficient method to route topologically a packet (in case of a failure) using an alternative path with no record at all of the original.
{"title":"Protection mechanism for the N2R topological routing algorithm","authors":"J. Lopez, J. Pedersen, R. C. Rumín, O. Madsen","doi":"10.1109/HSPR.2008.4734439","DOIUrl":"https://doi.org/10.1109/HSPR.2008.4734439","url":null,"abstract":"The topological routing over N2R structures has previously been studied and implemented using different techniques. A first approach was achieved obtaining the best trade off between path length vs. path completion time for the shortest path between any pair of nodes. This paper introduces protection against failures by modifying the previous algorithm implementing the option of routing a packet using a second independent path. The goal is to prove that there is an easy and efficient method to route topologically a packet (in case of a failure) using an alternative path with no record at all of the original.","PeriodicalId":130484,"journal":{"name":"2008 International Conference on High Performance Switching and Routing","volume":"262 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123045275","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：