is a new puradigm for iiano-scule communication between biological and artificial nalomachincs, i... more is a new puradigm for iiano-scule communication between biological and artificial nalomachincs, i,e,, small cemponents urid devices {such as proteins Emd ceJls) that pertbrm eemputation, sensing, or actuatien,
Leading the formulation of the SEI's technology strategy for R&D projects and external relationsh... more Leading the formulation of the SEI's technology strategy for R&D projects and external relationships by aligning the expertise of the SEI technical staff to identify and respond to the needs of sponsors, customers, and partners and help the SEI shape future innovations in complex softwarereliant systems.
The C++ language currently does not provide a mechanism for an object to determine its type at ru... more The C++ language currently does not provide a mechanism for an object to determine its type at runtime. We propose the Dossier class as a standard interface for accessing type information from within a C++ program. We have implemented a tool called mkdossier that automatically generates type information in a form that can be compiled and linked with an application. In the prototype implementation, a class must have a virtual function to access an object's dossier given the object. We propose this access be provided implicitly by the language through a predefined member in all classes.
International Journal of Cooperative Information Systems, Sep 1, 2006
This paper presents a semantics-based dynamic service composition architecture that composes an a... more This paper presents a semantics-based dynamic service composition architecture that composes an application through combining distributed components based on the semantics of the components. This architecture consists of a component model called Component Service Model with Semantics (CoSMoS), a middleware called Component Runtime Environment (CoRE), and a service composition mechanism called Semantic Graph based Service Composition (SeGSeC). CoSMoS represents the semantics of components. CoRE provides interfaces to discover and access components modeled by CoSMoS. SeGSeC composes an application by discovering components through CoRE, and synthesizing a workflow of the application based on the semantics of the components modeled by CoSMoS. This paper describes the latest design of the semantics-based dynamic service composition architecture, and also illustrates the implementation of the architecture based on the Web Service standards, i.e., WSDL, RDF, SOAP, and UDDI. The Web Service based implementation of the architecture allows existing Web Services to migrate onto the architecture without reimplementation. It also simplifies the development and deployment of a new Web Service on the architecture by automatically generating the necessary description files (i.e., WSDL and RDF files) of the Web Service from its runtime binary (i.e., a Java class file).
Packet collisions and their resolution create a performance bottleneck in random access LA\' s. A... more Packet collisions and their resolution create a performance bottleneck in random access LA\' s. A hardware solution to this problem is to u.~e collision avoidance switches [1, 2]. Collision avoidance switches allow the implementation of fandom access protocols without the penalty of collisions among packets. In this paper, we describe a design and implementation of a local area network ;:i rchi tecture based on collision avoidance, called the Collision Avoidance Multiple Broadcast (CA:\IB) tree netv.;ork. Our implementation includes CA.MB tree switches, station/network interface boards, and support of transport protocols. Our implementation of the CAME tree network follows the protocol layering architecture of the IEEE 802 local area networks.
To solve a performance bottle neck in random access LANs due to packet collisions and their resol... more To solve a performance bottle neck in random access LANs due to packet collisions and their resolution, collision avoidance switches are introduced. These switches allow random access protocols to achieve high performance by resolving collisions among packets. A conventional hardware implementation of these switches is the use of TTL chips. In this implementation; a handful of TTL chips are required to forma single switch (e.g., 18 TTL chips are needed for an implementation of the CAMB switch [7]). Thus, implementation of a complete network, which requires severa! of these switches, could very well result in a large and complex hardware system. Today's modern chip technology allows us to pack large quantity of logic in a single chip. By transferring the conventional implementation of the collision avoidance switches into a VLSI chip, the complexity of the resultant hardware is greatly reduced, not to mention the improvement in hardware performance and ease of packaging. This report provides an overall study of the collision avoidance protocols for the tree LANs with emphasis on the implementation of collision avoidance switches. Hardware implementations of sorne of these switches are discussed. And a VLSI implementation of the CAMB switch protocol is introduced.
The increasing demand for communication services, coupled with recent technological •advances in ... more The increasing demand for communication services, coupled with recent technological •advances in communication media and switching techniques, has resulted in a proliferation of new and expanded services. Currently, networks are needed which can transmit voice, data, and video services in an application-independent fashion. Unified approaches employ a single switching technique across the entire network bandwidth, thus, allowing services to be switched in an application-independent manner. This paper presents a taxonomy of integrated-service networks including a look at N-ISDN, while focusing on unified approaches to integrated-service networks. The two most promising unified approaches are burst and fast packet switching. Burst switching is a circuit switching-based approach which allocates channel bandwidth to a connection only during the transmission of "bursts" of information. Fast packet switching is a packet switching-based approach which can be characterized by very high transmission rates on network links and simple, hardwired protocols which match the rapid channel speed of the network. Both approaches are being proposed as possible implementations for integrated-service networks. We survey these two approaches, and also examine the key performance issues found in fast packet switching. We then present the results of a simulation study of a fast packet switching network.
Packet collisions and their resolution create a performance bottleneck in random-access LANs. A h... more Packet collisions and their resolution create a performance bottleneck in random-access LANs. A hardware solution to this problem is to use a collision avoidance switch. These switches allow the implementation of random access protocols without the penalty of collisions among packets. An architecture based on collision avoidance is the CAME (Collision Avoidance Multiple Broadcast) Tree network, where concurrent broadcasts are possible. The purpose of this paper is to present nvo implementations for a CAN!B Tree s~vitch•. First, a general outline of the CAME switch is provided. Then, a description of the nvo implementations is given.
We believe that the challenges faced by future network applications, such as scalability, adaptab... more We believe that the challenges faced by future network applications, such as scalability, adaptability, and survivability/availability have already been overcome by large scale biological systems and that future network applications will benefit by adopting key biological principles and mechanisms. The Bio-Networking Architecture is a paradigm as well as middleware that enables the construction and deployment of scalable, adaptive, and i survivable/available applications. In the Bio-Networking Architecture, applications are constructed using a collection of autonomous mobile agents, ralled cyber-entities. A cyber-entity is analogous to an individual bee in the biological world. Like their biological counterparts, cyber-entities follow biological principles and contain biological mechanisms. The desirable characteristics of an application emerge from the collective actions and interactions of its constituent cyber-entities. We describe a web content distribution application called Aphid, which was constructed using the Bio-Networking Architecture. Through simulations, we show that Aphid adapts to changing user demand and location. Aphid's scalability and survivability/availability are also demonstrated through simulations.
It is crucial for ATM to efficiently support existing data applications. One of the promising met... more It is crucial for ATM to efficiently support existing data applications. One of the promising methods of supporting data over ATM uses connectionless servers providing a packet switching service in the ATM network. There are two primary methods for forwarding cell based datagrams in a connectionless server: reassembly mode, where the datagrams are reconstructed before forwarding, and streaming mode, where each cell is forwarded as it arrives. The two modes impact the packet loss probability at a connectionless server, the packet delay, the interleaving method used, the AAL used, and the buffer space required. This paper investigates the performance of two forwarding modes. Extensive simulations show that forwarding connectionless packets in streaming mode is shown to result in lower packet loss and end-to-end delay than reassembly mode.
This paper describes a new architecture for high performance distributed applications and a suppo... more This paper describes a new architecture for high performance distributed applications and a supporting framework. This architecture applies object-oriented design and im plementation techniques to build a framework for platform-independent distributed ap plication specification and implementation using existing programming languages and operating systems. It utilizes an efficient and extensible layering architecture that al lows new abstract data types, new presentation-layer protocols, and new interprocess communication mechanisms to be added as they become necessary. Experimental re sults are presented demonstrating that the multiple layers of abstraction used do not compromise efficiency.
Layered transmission of data is often recommended as a solution to the problem of varying bandwid... more Layered transmission of data is often recommended as a solution to the problem of varying bandwidth constraints in multicast applications. In the case of video multicast, this technique encodes multiple interdependent layers of video at arbitrary target rates in order to address heterogeneous bandwidth constraints between the source and multiple receivers. However, multi-layered encoding alone is not sufficient to provide high video quality and high bandwidth utilization, because bandwidth constraints change over time. Adaptive techniques capable of adjusting the rates of video layers are required to maximize video quality and network utilization. In this paper we define a class of algorithms known as Source-Adaptive Multilayered Multicast (SAMM) algorithms. In SAMM algorithms, the source uses congestion feedback to adjust the number of generated layers and the bit rate of each layer. Furthermore, we introduce an end-to-end SAMM algorithm, in which only end systems monitor available bandwidth and report the amount of available bandwidth to the source. Using simulations which incorporate actual multi-layered video codecs, we demonstrate that the proposed SAMM algorithm exhibits better scalability and responsiveness to congestion than algorithms which are not source-adaptive.
This paper describes a transport system architecture called ADAPTIVE that we are developing to su... more This paper describes a transport system architecture called ADAPTIVE that we are developing to support multimedia applications utilizing high-speed networks. ADAPTIVE stands for "A Dynamically Assembled Protocol Transformation, Integration, and Validation Environment." It is a transformation-based system with the goal of providing policies and mechanisms for interactively specifying and configuring a flexible and adaptive transport system. In addition, it aims to provide a controlled prototyping environment for monitoring, analyzing, and experimenting with the effects of different transport system designs and implementations on application performance. The performance of distributed applications (such as remote terminal access, network file servers, distributed databases, collaborative work activities, bulk data transfer, computer imaging, audio-and video-conferencing, full-motion video, scientific computation and visualization) is influenced by network factors and transport system 1 factors. Several key network factors include channel speed, bit-error rate, and congestion levels at the intermediate switching nodes (ISNs). Transport system factors include both protocol processing activities (such as connection management, transmission control, and error protection) and general operating system hardware and software factors (such as memory latency, CPU speed, interrupt and context switching overhead, process architecture, and message buffering). Advances in fiber optics and VLSI technology have increased network channel speed by several orders of magnitude [1]. Coupled with the increase in application and network diversity described in Section 2, these changes are shifting performance bottlenecks from the network factors to the transport system factors. Therefore, transport system architectures must be more flexible and adaptive to handle the diversity and dynamism in network characteristics and application communication requirements [2]. However, existing transport systems, which typically offer only a small number of monolithic protocols, are unable to meet these flexible and adaptive requirements adequately [3]. This paper is organized as follows: in Section 2 we define four important research problems involving transport systems and describe how existing systems do not adequately solve these problems, Section 3 introduces the ADAPTIVE system and outlines its design, and Section 4 summarizes the paper and describes future work. 2 Research Problems Our research is investigating solutions to the following four problems involving transport systems: (A) The Throughput Preservation Problem: As noted by [1, 4, 5], only a limited fraction of the available bandwidth in high-speed networks is being delivered to applications. This situation results from transport system overhead not decreasing as rapidly as the network channel-speed is increasing. The overhead results from activities such as memory-to-memory copying and process management operations like interrupt handling, context switching, and scheduling. (B) The Multimedia Application Diversity and Dynamism Problem: Another issue (which is related to the throughput preservation problem) is how to support the communication requirements of distributed multimedia applications 2 that are more demanding and dynamic than those found in traditional data applications such as remote terminal access, file transfer, and electronic mail. For example, multimedia applications involve combinations of requirements such as extremely high throughput (e.g., full-motion video), strict realtime delivery (e.g., manufacturing control systems), low latency (e.g., on-line transaction processing), low delay jitter (e.g., voice conversation), multicast capability (e.g., collaborative work activities), high-reliability (e.g., bulk data transfer), temporal synchronization (e.g., audio-and video-conferencing), and some degree of loss tolerance (e.g., hierarchically-coded voice and video). Moreover, multimedia applications impose different network traffic patterns. For instance, some applications generate highly bursty traffic (e.g., variable bit-rate video applications), some generate continuous traffic (e.g., constant bit-rate video applications), and others generate short, interactive, transaction-oriented traffic (e.g., network file systems using remote Transport Example Average Burst Delay Jitter Order
The advent of high speed, wide area ATM networks introduces a "backwards compatibility" dilemma. ... more The advent of high speed, wide area ATM networks introduces a "backwards compatibility" dilemma. While emerging ATM networks must be capable of transporting future applications such as digital voice and video, many of the design choices made to optimize ATM for voice and video make the transport of traditional data traffic difficult. This paper will address some of these issues and propose possible solutions. This paper presents a promising method of providing connectionless service by building a virtual connectionless network on top of a public ATM network with a subset of the switches acting as con nectionless servers. This method provides low transfer delay and low cost to the user.
Fora variety ofreasons, distributed applications often must beimplemented using exist ing convent... more Fora variety ofreasons, distributed applications often must beimplemented using exist ing conventional programming languages and operating systems. Creating new high perfor mance distributed applications without high-level support from the programming language or operating system requires thatthe mechanism of distribution be selected early in the de sign stage, reducing the flexibility and/or efficiency of the design and subsequent imple mentation. Adding distribution to existing applications can result in an inordinate amount of reengineering duetothecomplexity andheterogeneity of interapplication communication mechanisms and their interfaces. This paper describes a new architecture for high performance distributed applications and a supporting framework. This architecture applies object-oriented design and imple mentation techniques tobuild a framework forplatform-independent distributed application specification and implementation using existing programming languages and operating sys tems. It utilizes an efficient and extensible layering architecture that allows new abstract data types, new protocols, and new interprocess communication mechanisms to be addedas they become necessary. h 6.
In an integrated service environment, where users exchange various types of aural and visual info... more In an integrated service environment, where users exchange various types of aural and visual information, networks should appear friendly to its users providing tools for management of multi-media information. Networks should also efficiently satisfy diverse performance requirements of different information being exchanged. In this paper we present new architecture for integrated service networks being investigated and developed by the Distributed Computation and Communication Group at the Department of Computer Science in the Columbia University. Research efforts are devoted to developing both (1) document management software to allow users to manipulate and relate text/graphics/voice information in a dynamic way, and (2) a tree network architecture for reliable and efficient exchange of multi-media information.
In IaaS cloud services, QoS of network applications (NW-Apps) may degrade due to location factors... more In IaaS cloud services, QoS of network applications (NW-Apps) may degrade due to location factors such as significant distance between a server-side application (server) of a NW-App at a data center and a client-side application (client) of the NW-App at a client terminal. In order to shorten the distance and to improve the QoS, server migration services (SMSes) have been proposed. In SMSes, servers may migrate between different computers (called work places, WPs) on a network to prevent QoS degradation caused by the changes of client locations. Although server migrations can improve QoS of NW-Apps, they also generate a huge amount of traffic (server migration traffic) in the network. This paper focuses on a server location decision problem where the location of a server is decided in an on-line manner so that QoS of a NW-App is improved under the constraint that the server migration traffic has to be suppressed below an acceptable level. For the problem, we propose a practical on-line algorithm. The key idea behind the proposed algorithm is that the location of the server is decided with consideration of the QoS degradation in the future. The algorithm defines the averagely good location for the server where the QoS is expected to be relatively good for various client locations. Then, it keeps the range of the server's migration within the returnable range where the server can soon come back to the averagely good location. As a result, the QoS can be always kept as good as the one under the averagely good location. Simulation results show that the proposed algorithm improves QoS of the NW-App by up to 30% compared to a greedy algorithm.
Tatsuya Suda (University of California at Irvine) ... Willis King (University of Houston) Mats Da... more Tatsuya Suda (University of California at Irvine) ... Willis King (University of Houston) Mats Daniels (University of Uppsala) Jane Prey (University of Virginia) Sally Fincher (University of Kent) Anne Parker (Eduprise) Gordon Davies (Open University) Stephen Seidman (Colorado State University) ... Workshop on Software Engineering and the Internet ... Patrick Bobbie (Southern Polytechnic State University) Jim Isaak (Independent Consultant) Ly Sauer (Sandia National Laboratories) Joseph Urban (Arizona State University)
is a new puradigm for iiano-scule communication between biological and artificial nalomachincs, i... more is a new puradigm for iiano-scule communication between biological and artificial nalomachincs, i,e,, small cemponents urid devices {such as proteins Emd ceJls) that pertbrm eemputation, sensing, or actuatien,
Leading the formulation of the SEI's technology strategy for R&D projects and external relationsh... more Leading the formulation of the SEI's technology strategy for R&D projects and external relationships by aligning the expertise of the SEI technical staff to identify and respond to the needs of sponsors, customers, and partners and help the SEI shape future innovations in complex softwarereliant systems.
The C++ language currently does not provide a mechanism for an object to determine its type at ru... more The C++ language currently does not provide a mechanism for an object to determine its type at runtime. We propose the Dossier class as a standard interface for accessing type information from within a C++ program. We have implemented a tool called mkdossier that automatically generates type information in a form that can be compiled and linked with an application. In the prototype implementation, a class must have a virtual function to access an object's dossier given the object. We propose this access be provided implicitly by the language through a predefined member in all classes.
International Journal of Cooperative Information Systems, Sep 1, 2006
This paper presents a semantics-based dynamic service composition architecture that composes an a... more This paper presents a semantics-based dynamic service composition architecture that composes an application through combining distributed components based on the semantics of the components. This architecture consists of a component model called Component Service Model with Semantics (CoSMoS), a middleware called Component Runtime Environment (CoRE), and a service composition mechanism called Semantic Graph based Service Composition (SeGSeC). CoSMoS represents the semantics of components. CoRE provides interfaces to discover and access components modeled by CoSMoS. SeGSeC composes an application by discovering components through CoRE, and synthesizing a workflow of the application based on the semantics of the components modeled by CoSMoS. This paper describes the latest design of the semantics-based dynamic service composition architecture, and also illustrates the implementation of the architecture based on the Web Service standards, i.e., WSDL, RDF, SOAP, and UDDI. The Web Service based implementation of the architecture allows existing Web Services to migrate onto the architecture without reimplementation. It also simplifies the development and deployment of a new Web Service on the architecture by automatically generating the necessary description files (i.e., WSDL and RDF files) of the Web Service from its runtime binary (i.e., a Java class file).
Packet collisions and their resolution create a performance bottleneck in random access LA\' s. A... more Packet collisions and their resolution create a performance bottleneck in random access LA\' s. A hardware solution to this problem is to u.~e collision avoidance switches [1, 2]. Collision avoidance switches allow the implementation of fandom access protocols without the penalty of collisions among packets. In this paper, we describe a design and implementation of a local area network ;:i rchi tecture based on collision avoidance, called the Collision Avoidance Multiple Broadcast (CA:\IB) tree netv.;ork. Our implementation includes CA.MB tree switches, station/network interface boards, and support of transport protocols. Our implementation of the CAME tree network follows the protocol layering architecture of the IEEE 802 local area networks.
To solve a performance bottle neck in random access LANs due to packet collisions and their resol... more To solve a performance bottle neck in random access LANs due to packet collisions and their resolution, collision avoidance switches are introduced. These switches allow random access protocols to achieve high performance by resolving collisions among packets. A conventional hardware implementation of these switches is the use of TTL chips. In this implementation; a handful of TTL chips are required to forma single switch (e.g., 18 TTL chips are needed for an implementation of the CAMB switch [7]). Thus, implementation of a complete network, which requires severa! of these switches, could very well result in a large and complex hardware system. Today's modern chip technology allows us to pack large quantity of logic in a single chip. By transferring the conventional implementation of the collision avoidance switches into a VLSI chip, the complexity of the resultant hardware is greatly reduced, not to mention the improvement in hardware performance and ease of packaging. This report provides an overall study of the collision avoidance protocols for the tree LANs with emphasis on the implementation of collision avoidance switches. Hardware implementations of sorne of these switches are discussed. And a VLSI implementation of the CAMB switch protocol is introduced.
The increasing demand for communication services, coupled with recent technological •advances in ... more The increasing demand for communication services, coupled with recent technological •advances in communication media and switching techniques, has resulted in a proliferation of new and expanded services. Currently, networks are needed which can transmit voice, data, and video services in an application-independent fashion. Unified approaches employ a single switching technique across the entire network bandwidth, thus, allowing services to be switched in an application-independent manner. This paper presents a taxonomy of integrated-service networks including a look at N-ISDN, while focusing on unified approaches to integrated-service networks. The two most promising unified approaches are burst and fast packet switching. Burst switching is a circuit switching-based approach which allocates channel bandwidth to a connection only during the transmission of "bursts" of information. Fast packet switching is a packet switching-based approach which can be characterized by very high transmission rates on network links and simple, hardwired protocols which match the rapid channel speed of the network. Both approaches are being proposed as possible implementations for integrated-service networks. We survey these two approaches, and also examine the key performance issues found in fast packet switching. We then present the results of a simulation study of a fast packet switching network.
Packet collisions and their resolution create a performance bottleneck in random-access LANs. A h... more Packet collisions and their resolution create a performance bottleneck in random-access LANs. A hardware solution to this problem is to use a collision avoidance switch. These switches allow the implementation of random access protocols without the penalty of collisions among packets. An architecture based on collision avoidance is the CAME (Collision Avoidance Multiple Broadcast) Tree network, where concurrent broadcasts are possible. The purpose of this paper is to present nvo implementations for a CAN!B Tree s~vitch•. First, a general outline of the CAME switch is provided. Then, a description of the nvo implementations is given.
We believe that the challenges faced by future network applications, such as scalability, adaptab... more We believe that the challenges faced by future network applications, such as scalability, adaptability, and survivability/availability have already been overcome by large scale biological systems and that future network applications will benefit by adopting key biological principles and mechanisms. The Bio-Networking Architecture is a paradigm as well as middleware that enables the construction and deployment of scalable, adaptive, and i survivable/available applications. In the Bio-Networking Architecture, applications are constructed using a collection of autonomous mobile agents, ralled cyber-entities. A cyber-entity is analogous to an individual bee in the biological world. Like their biological counterparts, cyber-entities follow biological principles and contain biological mechanisms. The desirable characteristics of an application emerge from the collective actions and interactions of its constituent cyber-entities. We describe a web content distribution application called Aphid, which was constructed using the Bio-Networking Architecture. Through simulations, we show that Aphid adapts to changing user demand and location. Aphid's scalability and survivability/availability are also demonstrated through simulations.
It is crucial for ATM to efficiently support existing data applications. One of the promising met... more It is crucial for ATM to efficiently support existing data applications. One of the promising methods of supporting data over ATM uses connectionless servers providing a packet switching service in the ATM network. There are two primary methods for forwarding cell based datagrams in a connectionless server: reassembly mode, where the datagrams are reconstructed before forwarding, and streaming mode, where each cell is forwarded as it arrives. The two modes impact the packet loss probability at a connectionless server, the packet delay, the interleaving method used, the AAL used, and the buffer space required. This paper investigates the performance of two forwarding modes. Extensive simulations show that forwarding connectionless packets in streaming mode is shown to result in lower packet loss and end-to-end delay than reassembly mode.
This paper describes a new architecture for high performance distributed applications and a suppo... more This paper describes a new architecture for high performance distributed applications and a supporting framework. This architecture applies object-oriented design and im plementation techniques to build a framework for platform-independent distributed ap plication specification and implementation using existing programming languages and operating systems. It utilizes an efficient and extensible layering architecture that al lows new abstract data types, new presentation-layer protocols, and new interprocess communication mechanisms to be added as they become necessary. Experimental re sults are presented demonstrating that the multiple layers of abstraction used do not compromise efficiency.
Layered transmission of data is often recommended as a solution to the problem of varying bandwid... more Layered transmission of data is often recommended as a solution to the problem of varying bandwidth constraints in multicast applications. In the case of video multicast, this technique encodes multiple interdependent layers of video at arbitrary target rates in order to address heterogeneous bandwidth constraints between the source and multiple receivers. However, multi-layered encoding alone is not sufficient to provide high video quality and high bandwidth utilization, because bandwidth constraints change over time. Adaptive techniques capable of adjusting the rates of video layers are required to maximize video quality and network utilization. In this paper we define a class of algorithms known as Source-Adaptive Multilayered Multicast (SAMM) algorithms. In SAMM algorithms, the source uses congestion feedback to adjust the number of generated layers and the bit rate of each layer. Furthermore, we introduce an end-to-end SAMM algorithm, in which only end systems monitor available bandwidth and report the amount of available bandwidth to the source. Using simulations which incorporate actual multi-layered video codecs, we demonstrate that the proposed SAMM algorithm exhibits better scalability and responsiveness to congestion than algorithms which are not source-adaptive.
This paper describes a transport system architecture called ADAPTIVE that we are developing to su... more This paper describes a transport system architecture called ADAPTIVE that we are developing to support multimedia applications utilizing high-speed networks. ADAPTIVE stands for "A Dynamically Assembled Protocol Transformation, Integration, and Validation Environment." It is a transformation-based system with the goal of providing policies and mechanisms for interactively specifying and configuring a flexible and adaptive transport system. In addition, it aims to provide a controlled prototyping environment for monitoring, analyzing, and experimenting with the effects of different transport system designs and implementations on application performance. The performance of distributed applications (such as remote terminal access, network file servers, distributed databases, collaborative work activities, bulk data transfer, computer imaging, audio-and video-conferencing, full-motion video, scientific computation and visualization) is influenced by network factors and transport system 1 factors. Several key network factors include channel speed, bit-error rate, and congestion levels at the intermediate switching nodes (ISNs). Transport system factors include both protocol processing activities (such as connection management, transmission control, and error protection) and general operating system hardware and software factors (such as memory latency, CPU speed, interrupt and context switching overhead, process architecture, and message buffering). Advances in fiber optics and VLSI technology have increased network channel speed by several orders of magnitude [1]. Coupled with the increase in application and network diversity described in Section 2, these changes are shifting performance bottlenecks from the network factors to the transport system factors. Therefore, transport system architectures must be more flexible and adaptive to handle the diversity and dynamism in network characteristics and application communication requirements [2]. However, existing transport systems, which typically offer only a small number of monolithic protocols, are unable to meet these flexible and adaptive requirements adequately [3]. This paper is organized as follows: in Section 2 we define four important research problems involving transport systems and describe how existing systems do not adequately solve these problems, Section 3 introduces the ADAPTIVE system and outlines its design, and Section 4 summarizes the paper and describes future work. 2 Research Problems Our research is investigating solutions to the following four problems involving transport systems: (A) The Throughput Preservation Problem: As noted by [1, 4, 5], only a limited fraction of the available bandwidth in high-speed networks is being delivered to applications. This situation results from transport system overhead not decreasing as rapidly as the network channel-speed is increasing. The overhead results from activities such as memory-to-memory copying and process management operations like interrupt handling, context switching, and scheduling. (B) The Multimedia Application Diversity and Dynamism Problem: Another issue (which is related to the throughput preservation problem) is how to support the communication requirements of distributed multimedia applications 2 that are more demanding and dynamic than those found in traditional data applications such as remote terminal access, file transfer, and electronic mail. For example, multimedia applications involve combinations of requirements such as extremely high throughput (e.g., full-motion video), strict realtime delivery (e.g., manufacturing control systems), low latency (e.g., on-line transaction processing), low delay jitter (e.g., voice conversation), multicast capability (e.g., collaborative work activities), high-reliability (e.g., bulk data transfer), temporal synchronization (e.g., audio-and video-conferencing), and some degree of loss tolerance (e.g., hierarchically-coded voice and video). Moreover, multimedia applications impose different network traffic patterns. For instance, some applications generate highly bursty traffic (e.g., variable bit-rate video applications), some generate continuous traffic (e.g., constant bit-rate video applications), and others generate short, interactive, transaction-oriented traffic (e.g., network file systems using remote Transport Example Average Burst Delay Jitter Order
The advent of high speed, wide area ATM networks introduces a "backwards compatibility" dilemma. ... more The advent of high speed, wide area ATM networks introduces a "backwards compatibility" dilemma. While emerging ATM networks must be capable of transporting future applications such as digital voice and video, many of the design choices made to optimize ATM for voice and video make the transport of traditional data traffic difficult. This paper will address some of these issues and propose possible solutions. This paper presents a promising method of providing connectionless service by building a virtual connectionless network on top of a public ATM network with a subset of the switches acting as con nectionless servers. This method provides low transfer delay and low cost to the user.
Fora variety ofreasons, distributed applications often must beimplemented using exist ing convent... more Fora variety ofreasons, distributed applications often must beimplemented using exist ing conventional programming languages and operating systems. Creating new high perfor mance distributed applications without high-level support from the programming language or operating system requires thatthe mechanism of distribution be selected early in the de sign stage, reducing the flexibility and/or efficiency of the design and subsequent imple mentation. Adding distribution to existing applications can result in an inordinate amount of reengineering duetothecomplexity andheterogeneity of interapplication communication mechanisms and their interfaces. This paper describes a new architecture for high performance distributed applications and a supporting framework. This architecture applies object-oriented design and imple mentation techniques tobuild a framework forplatform-independent distributed application specification and implementation using existing programming languages and operating sys tems. It utilizes an efficient and extensible layering architecture that allows new abstract data types, new protocols, and new interprocess communication mechanisms to be addedas they become necessary. h 6.
In an integrated service environment, where users exchange various types of aural and visual info... more In an integrated service environment, where users exchange various types of aural and visual information, networks should appear friendly to its users providing tools for management of multi-media information. Networks should also efficiently satisfy diverse performance requirements of different information being exchanged. In this paper we present new architecture for integrated service networks being investigated and developed by the Distributed Computation and Communication Group at the Department of Computer Science in the Columbia University. Research efforts are devoted to developing both (1) document management software to allow users to manipulate and relate text/graphics/voice information in a dynamic way, and (2) a tree network architecture for reliable and efficient exchange of multi-media information.
In IaaS cloud services, QoS of network applications (NW-Apps) may degrade due to location factors... more In IaaS cloud services, QoS of network applications (NW-Apps) may degrade due to location factors such as significant distance between a server-side application (server) of a NW-App at a data center and a client-side application (client) of the NW-App at a client terminal. In order to shorten the distance and to improve the QoS, server migration services (SMSes) have been proposed. In SMSes, servers may migrate between different computers (called work places, WPs) on a network to prevent QoS degradation caused by the changes of client locations. Although server migrations can improve QoS of NW-Apps, they also generate a huge amount of traffic (server migration traffic) in the network. This paper focuses on a server location decision problem where the location of a server is decided in an on-line manner so that QoS of a NW-App is improved under the constraint that the server migration traffic has to be suppressed below an acceptable level. For the problem, we propose a practical on-line algorithm. The key idea behind the proposed algorithm is that the location of the server is decided with consideration of the QoS degradation in the future. The algorithm defines the averagely good location for the server where the QoS is expected to be relatively good for various client locations. Then, it keeps the range of the server's migration within the returnable range where the server can soon come back to the averagely good location. As a result, the QoS can be always kept as good as the one under the averagely good location. Simulation results show that the proposed algorithm improves QoS of the NW-App by up to 30% compared to a greedy algorithm.
Tatsuya Suda (University of California at Irvine) ... Willis King (University of Houston) Mats Da... more Tatsuya Suda (University of California at Irvine) ... Willis King (University of Houston) Mats Daniels (University of Uppsala) Jane Prey (University of Virginia) Sally Fincher (University of Kent) Anne Parker (Eduprise) Gordon Davies (Open University) Stephen Seidman (Colorado State University) ... Workshop on Software Engineering and the Internet ... Patrick Bobbie (Southern Polytechnic State University) Jim Isaak (Independent Consultant) Ly Sauer (Sandia National Laboratories) Joseph Urban (Arizona State University)
Uploads
Papers by Tatsuya Suda