The ADAPTIVE Service eXecutive (ASX) is a highly modular and extensible object-oriented framework that simplifies the development and configuration of distributed applications on shared memory multi-processor platforms. This paper describes the structure and functionality of the ASX framework 's object-oriented architecture. In addition, the paper presents the results of performance experiments conducted using ASX-based implementations of connection-oriented and connectionless protocols from the TCP/IP protocol family. These experiments measure the performance impact of alternative methods for parallelizing communication protocol stacks. Throughout the paper, examples are presented to indicate how the use of object-oriented techniques facilitate application extensibility, component reuse, and performance enhancement. 1 Introduction Distributed computing is a promising technology for improving collaboration through connectivity and interworking; performance through parallel processing...

Parallel processing has been proposed as a means of improving network protocol throughput. Several different strategies have been taken towards parallelizing protocols. A relatively popular approach is packet-level parallelism, where packets are distributed across processors. This paper provides an experimental performance study of packet-level parallelism on a contemporary sharedmemory multiprocessor. We examine several unexplored areas in packet-level parallelism and investigate how various protocol structuring and implementation techniques can affect performance. We study TCP/IP and UDP/IP protocol stacks, implemented with a parallel version of the x-kernel running in user space on Silicon Graphics multiprocessors. Our results show that only limited packet-level parallelism can be achieved within a single connection under TCP, but that using multiple connections can improve available parallelism. We also demonstrate that packet ordering plays a key role in determining single-connection TCP performance, that careful use of locks is a necessity, and that selective exploitation of caching can improve throughput. We also describe experiments that compare parallel protocol performance on two generations of a parallel machine and show how computer architectural trends can influence performance. 1

Over the next several years the performance demands on globally available information servers are expected to increase dramatically. These servers must be capable of sending and receiving data over hundreds or even thousands of simultaneous connections. In this paper, we show that connection-level parallel protocols (where different connections are processed in parallel) running on a sharedmemory multiprocessor can deliver high network bandwidth across a large number of connections. We experimentally evaluate connection-level parallel implementations of both TCP/IP and UDP/IP protocol stacks. We focus on three questions in our performance evaluation: how throughput scales with the number of processors, how throughput changes as the number of connections increases, and how fairly the aggregate bandwidth is distributed across connections. We show how several factors impact performance: the number of processors used, the number of threads in the system, the number of connections assigned ...

A communication subsystem consists of protocol functions and operating system mechanisms that support the implementation and execution of protocol stacks. To effectively parallelize a communication subsystem, careful consideration must be given to the process architecture used to structure multiple processing elements. A process architecture binds one or more processing elements with the protocol tasks and messages associated with protocol stacks in a communication subsystem. This paper outlines the two fundamental types of process architectures (task-based and messagebased) and describes performance experiments conducted on three representative examples of these two types of process architectures -- Layer Parallelism, which is a task-based process architecture, and Message-Parallelism and Connectional Parallelism, which are message-based process architectures. These experiments measure the impact of the process architecture on connectionless and connection-oriented protocol stacks (based upon UDP and TCP) in a sharedmemory multi-processor operating system. The results from these experiments indicate that the choice of process architecture significantly affects communication subsystem performance. 1

A communication subsystem consists of protocol tasks and operating system mechanisms that support the configuration and execution of protocol stacks composed of protocol tasks. To parallelize a communication subsystem effectively, careful consideration must be given to the threading architecture. The threading architecture binds processing elements with the protocol tasks and the messages associated with protocol stacks in a communication subsystem. This paper makes two contributions to the study and application of threading architectures. First, it reports performance results from empirical comparisons of two protocol stacks (based on the connectionless and connection-oriented transport protocols UDP and TCP) using different threading architectures on a 20 CPU multi-processor platform. The results demonstrate how andwhy different threading architectures affect performance. Second, the paper provides guidelines based on these results that indicate when and how to apply appropriate thre...

Security and privacy are growing concerns in the Internet community, due to the Internet's rapid growth and the desire to conduct business over it safely. This desire has led to the advent of several proposals for security standards, such as secure IP, secure HTTP, and the Secure Socket Layer. All of these standards propose using cryptographic protocols such as DES and RSA. Thus, the need to use encryption protocols is increasing. Shared-memory multiprocessors make attractive server platforms, for example as secure World-Wide Web servers. These machines are becoming more common, as shown by recent vendor introductions of platforms such as SGI's Challenge, Sun's SPARCCenter, and DEC's AlphaServer. The spread of these machines is due both to their relative ease of programming and their good price/performance. This paper is an experimental performance study that examines how encryption protocol performance can be improved by using parallelism. We show linear speedup for several different ...

There are many differences between ATM and todays most commonly used network technologies. New firewall architectures are required to exploit the advantages of ATM technology and to support the high throughput available in ATM networks. This paper begins with a discussion of the impact of ATM on firewalls and then introduces the idea of parallelized firewalls, which may be used in order to achieve the high performance necessary for ATM networks. 1 Introduction Firewalls are a widely used security mechanism in the Internet today. They are mostly used to provide access control and audit at the border between the public Internet and private networks, but are also used to secure critical subnets within private networks. ATM is another somewhat newer trend in networking today. ATM provides a scalable high-speed network infrastructure, based on the concepts of fixedlength cells and virtual circuits. These conceptual differences to "legacy" networks 1 and the high throughput of ATM netw...

Providing end-to-end gigabit communication support for bandwidth-intensive distributed applications requires highperformance transport systems. This paper describes and classifies transport system mechanisms that integrate operating system resources (such as CPU(s), virtual memory, and network adapters) together with communication protocols (such as TCP/IP and XTP) to support applications running on local and wide area networks. A taxonomy is presented that compares and evaluates four widely available transport systems in terms of their support for protocol processing. The systems covered in this paper include System V UNIX STREAMS, the BSD UNIX networking subsystem, the x-kernel, and the Conduit framework from the Choices operating system. This paper is intended to help researchers navigate through the transport system design space by describing alternative mechanisms for developing transport systems. 1 Introduction The demand for many types of distributed applications is expanding r...

The ADAPTIVE Service eXecutive (ASX) is an objectoriented framework that enhance the development of distributed applications across a range of operating system platforms. The components in ASX were developed using object-oriented design techniques and C++ language features in order to simplify the use of OS mechanisms that provide interprocess communication, communication port demultiplexing, explicit dynamic linking, and concurrency. In addition, the ASX components automate many system configuration and reconfiguration steps by dynamically linking network services into applications at run-time and arranging to execute these services on one or more processes or threads. This paper describes the structure and functionality of ASX and presents several examples illustrating key ASX features. 1 Introduction Developing communication systems that effectively utilize multi-processing and network services is a promising technique for increasing system performance, scalability, and cost effec...

We describe a framework for composing end-to-end protocol functions. The framework comprises: a generic model of protocol processing; a metaheader protocol supporting per-packet configuration of protocol function and efficient demultiplexing of incoming data units; and an extensible set of modular protocol functions. This paper describes the pieces of the framework and motivates some of the design decisions.