To enhance performance characteristics of data transport over ATM-UBR service class, additional mechanisms, such as congestion control and buffer management, are required. In this paper, we investigate active queue management mechanisms that are designed to enable ATM switches to operate in the low delay/high throughput regime. In particular, we present an analytical framework for random early detection (RED)-like congestion control mechanisms in an ATM multiplexer, and obtain closed form expressions for performance metrics such as throughput and average queueing delays. Based on this analysis, we propose an adaptive mechanism to select the parameters of the congestion control algorithm to optimize the system performance.
{"title":"A random early discard framework for congestion control in ATM networks","authors":"A. Mokhtar, M. Azizoglu","doi":"10.1109/ATM.1999.786776","DOIUrl":"https://doi.org/10.1109/ATM.1999.786776","url":null,"abstract":"To enhance performance characteristics of data transport over ATM-UBR service class, additional mechanisms, such as congestion control and buffer management, are required. In this paper, we investigate active queue management mechanisms that are designed to enable ATM switches to operate in the low delay/high throughput regime. In particular, we present an analytical framework for random early detection (RED)-like congestion control mechanisms in an ATM multiplexer, and obtain closed form expressions for performance metrics such as throughput and average queueing delays. Based on this analysis, we propose an adaptive mechanism to select the parameters of the congestion control algorithm to optimize the system performance.","PeriodicalId":266412,"journal":{"name":"IEEE ATM Workshop '99 Proceedings (Cat. No. 99TH8462)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1999-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121333914","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}
This paper considers a problem that arises in the context of statistical, state-dependent resource allocation in communication networks. Specifically, the problem concerns a single-server, discrete-time queue with state-dependent probabilistic server availability/interruptions, and is motivated by the analysis of ground station queueing dynamics in a broadband satellite environment, as well as the analysis of a blocking shared-medium ATM switch with input queueing. The system is analyzed for two forms of probabilistic behavior: exponential and hyperbolic forms. The analysis yields several performance metrics of interest, including average system occupancy, average delay, and steady-state occupancy distribution.
{"title":"On the performance of discrete-time queues with probabilistic server availability","authors":"A. Mokhtar, M. Azizoglu","doi":"10.1109/ATM.1999.786900","DOIUrl":"https://doi.org/10.1109/ATM.1999.786900","url":null,"abstract":"This paper considers a problem that arises in the context of statistical, state-dependent resource allocation in communication networks. Specifically, the problem concerns a single-server, discrete-time queue with state-dependent probabilistic server availability/interruptions, and is motivated by the analysis of ground station queueing dynamics in a broadband satellite environment, as well as the analysis of a blocking shared-medium ATM switch with input queueing. The system is analyzed for two forms of probabilistic behavior: exponential and hyperbolic forms. The analysis yields several performance metrics of interest, including average system occupancy, average delay, and steady-state occupancy distribution.","PeriodicalId":266412,"journal":{"name":"IEEE ATM Workshop '99 Proceedings (Cat. No. 99TH8462)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1999-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122521288","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}
We propose a new ATM switch with flexible buffering architecture and priority scheduling for providing a variety of QoS classes in this paper. In addition, this paper is concerned with the problem of supporting multiple QoS classes according to cell loss sensitivity and cell delay sensitivity. Loss-sensitive cells are given a higher priority to access a buffer with large free space to ensure a smaller cell loss probability, while delay-sensitive cells are assigned to a smaller buffer with high access priority to the output link to meet the shorter delay requirements. We also describe the implementation of a flexible ATM switch, FLASH (flexible ATM switch with high performance). FLASH also allows for multi-rate switching, supporting arbitrary-sized logical bit pipes with varying bandwidths. A notable characteristic is that FLASH provides multi-channel switching and guarantees cell sequence integrity without additional resequencing logic in the output port processors.
{"title":"Design of FLASH switch with flexible controllability","authors":"K. Kim, B. Ko, Jae Geun Kim, J. Choi","doi":"10.1109/ATM.1999.786876","DOIUrl":"https://doi.org/10.1109/ATM.1999.786876","url":null,"abstract":"We propose a new ATM switch with flexible buffering architecture and priority scheduling for providing a variety of QoS classes in this paper. In addition, this paper is concerned with the problem of supporting multiple QoS classes according to cell loss sensitivity and cell delay sensitivity. Loss-sensitive cells are given a higher priority to access a buffer with large free space to ensure a smaller cell loss probability, while delay-sensitive cells are assigned to a smaller buffer with high access priority to the output link to meet the shorter delay requirements. We also describe the implementation of a flexible ATM switch, FLASH (flexible ATM switch with high performance). FLASH also allows for multi-rate switching, supporting arbitrary-sized logical bit pipes with varying bandwidths. A notable characteristic is that FLASH provides multi-channel switching and guarantees cell sequence integrity without additional resequencing logic in the output port processors.","PeriodicalId":266412,"journal":{"name":"IEEE ATM Workshop '99 Proceedings (Cat. No. 99TH8462)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1999-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122979961","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}
A. Banchs, M. Gabrysch, T. Dietz, B. Lange, H. Stuttgen
In this paper different approaches to support multicast in ADSL networks are discussed and compared in terms of scalability, performance and implementation effort. Existing approaches, such as the multicast gateway or RAS replication, have the advantage of easy implementation, but suffer from undesirable scaling properties and low performance. On the other hand, more scalable solutions, such as MARS/MCS, rely on SVCs, which are currently not available in ADSL-based broadband networks. Thus, so far there is no solution based on currently available PVC operation providing a good performance. In order to cover this need, we propose a new architecture: multicast over PVC. This architecture, besides providing good performance, gives the possibility of an easy migration towards future SVC operation.
{"title":"Supporting multicast in ADSL networks","authors":"A. Banchs, M. Gabrysch, T. Dietz, B. Lange, H. Stuttgen","doi":"10.1109/ATM.1999.786872","DOIUrl":"https://doi.org/10.1109/ATM.1999.786872","url":null,"abstract":"In this paper different approaches to support multicast in ADSL networks are discussed and compared in terms of scalability, performance and implementation effort. Existing approaches, such as the multicast gateway or RAS replication, have the advantage of easy implementation, but suffer from undesirable scaling properties and low performance. On the other hand, more scalable solutions, such as MARS/MCS, rely on SVCs, which are currently not available in ADSL-based broadband networks. Thus, so far there is no solution based on currently available PVC operation providing a good performance. In order to cover this need, we propose a new architecture: multicast over PVC. This architecture, besides providing good performance, gives the possibility of an easy migration towards future SVC operation.","PeriodicalId":266412,"journal":{"name":"IEEE ATM Workshop '99 Proceedings (Cat. No. 99TH8462)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1999-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116250564","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}
S. Yoshikawa, H. Nagano, T. Suzuki, H. Nakane, M. Shinohara, T. Murase, G. Ramamurthy
ATM switches, which serve as nodes in ATM public backbone networks, are required to have a large capacity due to the increase of traffic. This paper proposes a large-scale ATM switch architecture providing 160 Gbit/s switching capability. It is such that the output buffered switch elements are expanded in square grids. Our switch architecture achieves non-blocking and multi-QoS guarantee. As a way to ensure multi-QoS, we employ a stop-shape-go (SSG) congestion control method and its performance is evaluated by simulation.
{"title":"Large scale input and output buffered ATM switch","authors":"S. Yoshikawa, H. Nagano, T. Suzuki, H. Nakane, M. Shinohara, T. Murase, G. Ramamurthy","doi":"10.1109/ATM.1999.786847","DOIUrl":"https://doi.org/10.1109/ATM.1999.786847","url":null,"abstract":"ATM switches, which serve as nodes in ATM public backbone networks, are required to have a large capacity due to the increase of traffic. This paper proposes a large-scale ATM switch architecture providing 160 Gbit/s switching capability. It is such that the output buffered switch elements are expanded in square grids. Our switch architecture achieves non-blocking and multi-QoS guarantee. As a way to ensure multi-QoS, we employ a stop-shape-go (SSG) congestion control method and its performance is evaluated by simulation.","PeriodicalId":266412,"journal":{"name":"IEEE ATM Workshop '99 Proceedings (Cat. No. 99TH8462)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1999-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125654344","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}
This paper presents experiments carried out with a TINA-based service, in the area of ATM connection management. Conclusions are reported on the flexibility of the service, with respect to its modularity, ease of use and deployment, as well as on its performance and scalability characteristics.
{"title":"For a flexible and scaleable ATM connection management: experiments","authors":"N. Charton","doi":"10.1109/ATM.1999.786895","DOIUrl":"https://doi.org/10.1109/ATM.1999.786895","url":null,"abstract":"This paper presents experiments carried out with a TINA-based service, in the area of ATM connection management. Conclusions are reported on the flexibility of the service, with respect to its modularity, ease of use and deployment, as well as on its performance and scalability characteristics.","PeriodicalId":266412,"journal":{"name":"IEEE ATM Workshop '99 Proceedings (Cat. No. 99TH8462)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1999-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132564961","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}
ATM adaptation layer type 2 (AAL2) has been standardized to carry highly compressed voice traffic with silence suppression over an ATM network. The ITU-T recommends using the constant bit rate (CBR) for ATM service carried through an AAL2 virtual channel (VC) to guarantee voice quality. However, the traffic characteristic of AAL2 VC is variable bit rate (VBR), which is different from that of the TCP/IP data traffic at the ATM cell level by two multiplexing techniques: priority multiplexing and non-priority multiplexing. Priority multiplexing enables the transfer to TCP/IP data with the AAL2 voice quality guaranteed. Non-priority multiplexing can enable the application of VBR services through AAL2 under certain conditions, and can thus help optimize transmission efficiency in the ATM backbone network.
ATM适配层类型2 (AAL2)已被标准化,以在ATM网络上承载高度压缩的语音业务,并具有沉默抑制。ITU-T建议通过AAL2虚拟信道(VC)进行ATM业务时使用CBR (constant bit rate),以保证语音质量。然而,AAL2 VC的流量特性是可变比特率(VBR),它通过优先级复用和非优先级复用两种复用技术区别于ATM单元级的TCP/IP数据流量。优先级复用可以在保证AAL2语音质量的情况下传输到TCP/IP。非优先级复用可以在一定条件下通过AAL2实现VBR业务的应用,从而优化ATM骨干网的传输效率。
{"title":"Performance evaluation of multiplexing AAL2 voice traffic and TCP/IP data at the ATM cell level","authors":"T. Okutani, H. Watanabe, T. Nisase","doi":"10.1109/ATM.1999.786887","DOIUrl":"https://doi.org/10.1109/ATM.1999.786887","url":null,"abstract":"ATM adaptation layer type 2 (AAL2) has been standardized to carry highly compressed voice traffic with silence suppression over an ATM network. The ITU-T recommends using the constant bit rate (CBR) for ATM service carried through an AAL2 virtual channel (VC) to guarantee voice quality. However, the traffic characteristic of AAL2 VC is variable bit rate (VBR), which is different from that of the TCP/IP data traffic at the ATM cell level by two multiplexing techniques: priority multiplexing and non-priority multiplexing. Priority multiplexing enables the transfer to TCP/IP data with the AAL2 voice quality guaranteed. Non-priority multiplexing can enable the application of VBR services through AAL2 under certain conditions, and can thus help optimize transmission efficiency in the ATM backbone network.","PeriodicalId":266412,"journal":{"name":"IEEE ATM Workshop '99 Proceedings (Cat. No. 99TH8462)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1999-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"113995674","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}
An increase in high quality of service (QoS) applications such as video conferencing and distribution, and the evolution of the Internet have popularized ATM-LAN use based on the private network node interface (PNNI). Also in public networks, which serve as backbone networks for LANs, ATM technology is being introduced for high-speed and broadband communication. These situations lead to a great demand for economically and seamlessly interconnecting remote ATM-LANs via ATM public networks, which are based on the broadband ISDN user part (BISUP). This paper discusses a peer group configuration method for such an internetworking architecture that can avoid an overload of the PNNI routing processing in each peer group. The paper also proposes a method for seamless interconnection of remote ATM-LANs. In this method, complete PNNI signalling and routing is executed between a local switch (LS) in a public network and each ATM-LAN. It can also reduce the PNNI routing processing load on each public network by emulating PNNI routing and signalling between LSs.
{"title":"PNNI internetworking architecture over ATM public networks","authors":"M. Kakemizu, K. Murata, M. Wakamoto","doi":"10.1109/ATM.1999.786881","DOIUrl":"https://doi.org/10.1109/ATM.1999.786881","url":null,"abstract":"An increase in high quality of service (QoS) applications such as video conferencing and distribution, and the evolution of the Internet have popularized ATM-LAN use based on the private network node interface (PNNI). Also in public networks, which serve as backbone networks for LANs, ATM technology is being introduced for high-speed and broadband communication. These situations lead to a great demand for economically and seamlessly interconnecting remote ATM-LANs via ATM public networks, which are based on the broadband ISDN user part (BISUP). This paper discusses a peer group configuration method for such an internetworking architecture that can avoid an overload of the PNNI routing processing in each peer group. The paper also proposes a method for seamless interconnection of remote ATM-LANs. In this method, complete PNNI signalling and routing is executed between a local switch (LS) in a public network and each ATM-LAN. It can also reduce the PNNI routing processing load on each public network by emulating PNNI routing and signalling between LSs.","PeriodicalId":266412,"journal":{"name":"IEEE ATM Workshop '99 Proceedings (Cat. No. 99TH8462)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1999-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116113198","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}
This paper proposes simple architecture to support multicasting in ATM switches. The proposed architecture consists of a self-routing switch which enables us to copy cell and line interfaces with a header conversion part. Our finding is that multicast can be realized by the method as follows; (1) copying cells in the switch; (2) recurring the copied cells from output ports to input ports; (3) changing the header information; (4) inputting the cells to the switch again. This proposed recursive multicast method has four main advantages. First, we can use normal line interfaces for the multicast function. Secondly, the number of lines connecting input and output can change the performance of the multicast very easily. Thirdly, this method can lower the cost of the hardware more than other approaches. Finally, it enables us to input plural cells in one physical port. In addition, this method has proved to provide an acceptable quality after experiment using a paper prototype system.
{"title":"Multicast ATM switch architecture based on a recursive method","authors":"M. Wada, K. Nakayama, K. Sakamoto","doi":"10.1109/ATM.1999.786899","DOIUrl":"https://doi.org/10.1109/ATM.1999.786899","url":null,"abstract":"This paper proposes simple architecture to support multicasting in ATM switches. The proposed architecture consists of a self-routing switch which enables us to copy cell and line interfaces with a header conversion part. Our finding is that multicast can be realized by the method as follows; (1) copying cells in the switch; (2) recurring the copied cells from output ports to input ports; (3) changing the header information; (4) inputting the cells to the switch again. This proposed recursive multicast method has four main advantages. First, we can use normal line interfaces for the multicast function. Secondly, the number of lines connecting input and output can change the performance of the multicast very easily. Thirdly, this method can lower the cost of the hardware more than other approaches. Finally, it enables us to input plural cells in one physical port. In addition, this method has proved to provide an acceptable quality after experiment using a paper prototype system.","PeriodicalId":266412,"journal":{"name":"IEEE ATM Workshop '99 Proceedings (Cat. No. 99TH8462)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1999-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123568954","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}
This paper presents an adaptation of the ERICA and ERICA+ algorithms for HFC (hybrid fibre coaxial) networks. It is derived analytically, and confirmed by simulations, that the admission rate at which new sources can begin to transmit is a time independent constant, that depends only on the total available bandwidth and on the number of data cells between two consecutive RM cells. For low bandwidth values, this admission rate implies high access latencies. This occurs if we rely only on the congestion avoidance and fairness achievement computations of ERICA. A new procedure is proposed in order to cope with this limitation, which reduces substantially the access latencies and the maximum burst transfer delay. Simulations results both with continuous and with bursty sources are presented. A study and simulation of interactive bursty traffic is also presented.
{"title":"Adaptation of the ERICA algorithm to HFC networks","authors":"M. Macedo, M. Nunes, H. Duarte-Ramos","doi":"10.1109/ATM.1999.786853","DOIUrl":"https://doi.org/10.1109/ATM.1999.786853","url":null,"abstract":"This paper presents an adaptation of the ERICA and ERICA+ algorithms for HFC (hybrid fibre coaxial) networks. It is derived analytically, and confirmed by simulations, that the admission rate at which new sources can begin to transmit is a time independent constant, that depends only on the total available bandwidth and on the number of data cells between two consecutive RM cells. For low bandwidth values, this admission rate implies high access latencies. This occurs if we rely only on the congestion avoidance and fairness achievement computations of ERICA. A new procedure is proposed in order to cope with this limitation, which reduces substantially the access latencies and the maximum burst transfer delay. Simulations results both with continuous and with bursty sources are presented. A study and simulation of interactive bursty traffic is also presented.","PeriodicalId":266412,"journal":{"name":"IEEE ATM Workshop '99 Proceedings (Cat. No. 99TH8462)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1999-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129147660","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}