This paper describes a new operating system kernel, called the x-kernel, that provides an explicit architecture for constructing and composing network protocols. Our experience implementing and evaluating several protocols in the x-kernel shows that this architecture is both general enough to accommodate a wide range of protocols, yet efficient enough to perform competitively with less structured operating systems. 1 Introduction Network software is at the heart of any distributed system. It manages the communication hardware that connects the processors in the system and it defines abstractions through which processes running on those processors exchange messages. Network software is extremely complex: it must hide the details of the underlying hardware, recover from transmission failures, ensure that messages are delivered to the application processes in the appropriate order, and manage the encoding and decoding of data. To help manage this complexity, network software is divi...

This paper presents the design, implementation and performance of a reliable multicast transport protocol called RMTP. RMTP is based on a hierarchical structure in which receivers are grouped into local regions or domains and in each domain there is a special receiver called a Designated Receiver (DR) which is responsible for sending acknowledgments periodically to the sender, for processing acknowledgements from receivers in its domain and for retransmitting lost packets to the corresponding receivers. Since lost packets are recovered by local retransmissions as opposed to retransmissions from the original sender, end-to-end latency is significantly reduced, and the overall throughput is improved as well. Also, since only the DRs send their acknowledgments to the sender, instead of all receivers sending their acknowledgments to the sender, a single acknowledgement is generated per local region, and this prevents acknowledgement implosion. Receivers in RMTP send their acknowledgments to the DRs periodically, thereby simplifying error recovery. In addition, lost packets are recovered by selective repeat retransmissions, leading to improved throughput at the cost of minimal additional buffering at the receivers. This paper also describes the implementation of RMTP and its performance on the Internet.

ion Psync is based on a conversation abstraction that provides a shared message space through which a collection of processes exchange messages. The general form of this message space is defined by a directed acyclic graph that preserves the partial order of the exchanged messages. For the purpose of this section, we view a conversation as an abstract data type that is implemented in shared memory; Section 3 gives an algorithm for implementing a conversation in an unreliable network. A conversation behaves much like any connection-oriented IPC abstraction: A well-defined set of processes---called participants---explicitly open a conversation, exchange messages through it, and close the conversation. Only processes that have been identified as participants may exchange message through the conversation, and this set is fixed for the duration of the conversation. Processes begin a conversation with the operations: conv = active open(participant set) conv = passive open(pid) The first...

Abstract — We study the role of pricing policies in multiple service class networks. We first argue that some form of serviceclass sensitive pricing is required for any multiclass service discipline ’ to attain the desired level of performance. Borrowing heavily from the Nash implementation paradigm in economics, we then present an abstract formulation of service disciplines and pricing policies. This formulation allows us to describe more clearly the interplay between service disciplines and pricing policies in determining overall network performance. Effective mrdtichtss service disciplines allow networks to focus resources on performance sensitive applications, while effective pricing policies allow us to spread the benefits of multiple service classes around to all users, rather than just having these benefits remain exclusively with the users of applications that are performance sensitive. Furthermore, service disciplines and pricing policies combine to form the incentive system facing a user; these incentives must be carefully tuned so that user self-interest leads to optimaf overall network performance. Finally, we illustrate some of these concepts through simulation of several simple example networks. In our simulations, we find that it is possible to set the prices so that users of every application type are more satisfied with the combined cost and performance of a network with service-class sensitive prices. For some application types the performance penalty received for requesting a less-than-optimal service class is offset by the reduced price of the service. For the other application types the monetary penalty incurred by using the more expensive, higher-quality service classes is offset by the improved performance they receive.z I.

Network software is a critical component of any distributed system. Because of its complexity, network software is commonly layered into a hierarchy of protocols, or more generally, into a protocol graph Typical protocol graphs-including those standardized in the IS0 and TCP/IP network architectures-share three important properties: the protocol graph is simple, the nodes of the graph (protocols) encapsulate complex functionality, and the topology of the graph is relatively static. This paper describes a new way to organize network software that differs from conventional architectures in all three of these properties In our approach, the protocol graph is complex, individual protocols encapsulate a single function. and the topology of the graph is dynamic. The main contribution of this paper is to describe the ideas behind our new architec-ture, illustrate the advantages of using the architecture, and demonstrate that the architecture results in efficient network software.

This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress". The list of current Internet-Drafts can be accessed at

The Internet Engineering Task Force (IETF) is in the process of adopting standards for IP-layer encryption and authentication (IPSEC). We describe a number of attacks against various versions of these protocols, including confidentiality failures and authentication failures. The implications of these attacks are troubling for the utility of this entire effort. 1 Introduction The Internet Engineering Task Force (IETF) is in the process of adopting standards for IP-layer encryption and authentication (IPSEC) [Atk95c, Atk95a, Atk95b, MS95, MKS95a]. While these protocols should provide a marked increase in Internet security, they themselves have had a checkered history. It is very much worth recounting the design history, not just to avoid the "oral history" problem in the IPSEC working group, but also because we as a profession learn more from knowing what doesn't work. As a wise sage 1 once said, "Learn from the mistakes of others; you'll never live long enough to make them all yours...

Modern switched networks such as ATM and Myrinet enable low-latency, high-bandwidth communication. This performance has not been realized by current applications, because of the high processing overheads imposed by existing communications software. These overheads are usually not hidden with large packets; most network traffic is small. We have developed Fast Sockets, a local-area communication layer that utilizes a high-performance protocol and exports the Berkeley Sockets programming interface. Fast Sockets realizes round-trip transfer times of 60 microseconds and maximum transfer bandwidth of 33 MB/second between two UltraSPARC 1s connected by a Myrinet network. Fast Sockets obtains performance by collapsing protocol layers, using simple buffer management strategies, and utilizing knowledge of packet destinations for direct transfer into user buffers. Using receive posting, we make the Sockets API a single-copy communications layer and enable regular Sockets programs to exploit the performance of modern networks. Fast Sockets transparently reverts to standard TCP/IP protocols for wide-area communication.

We study the role of pricing policies in multiple service class networks. We argue that some form of graduated prices are required in order for any multiclass service discipline to have the desired effect. Moreover, we demonstrate through simulation that it is possible to set the prices so that every user is more satisfied with the combined cost and performance of a network with graduated prices. For some users the performance penalty received for requesting a less-than-optimal service class is offset by the reduced price of the service. For the other users the monetary penalty incurred by using the more expensive, higher quality service classes is offset by the improved performance they receive. Thus, prices allow us to spread the benefits of multiple service classes around to all users, rather than just having these benefits remain exclusively with users who are performance sensitive. 1 Introduction Recent research on computer networks has been concerned almost exclusively with the...

This paper reports a series of experiments to measure TCP performance when transferring data through an Asynchronous Transfer Mode (ATM) switch. The results show that TCP buffer sizes and the ATM interface maximum transmission unit have a dramatic impact on throughput. We observe a throughput anomaly in which an increase in the receiver's buffer size decreases throughput substantially. For example, when using a 16K octet send buffer and ATM Adaptation Layer 5 on a 100 megabit per second (Mb/s) ATM path, the mean throughput for a bulk transfer drops from 15.05 Mb/s to 0.322 Mb/s if the receiver's buffer size is increased from 16K octets to 24K octets. This paper analyzes the performance, explains the anomalous behavior, and describes a solution that prevent the anomaly from occurring. This work was supported in part by a fellowship from UniForum Association. 1 Introduction Asynchronous Transfer Mode (or ATM) is a connection-oriented data communication technology that switches 53-octe...