A High Performance Communication Subsystem for PODOS

نویسندگان

  • Sudharshan S. Vazhkudai
  • P. Tobin Maginnis
چکیده

PODOS is a performance oriented distributed operating system being developed to harness the performance capabilities of a cluster computing environment. In order to address the growing demand for performance, we are designing a Distributed Operating System (DOS) that can utilize the computing potential of a number of systems. Earlier clustering approaches have traditionally stressed more on resource sharing or reliability and have given lesser priority to performance. PODOS adds just four new components to the existing Linux operating system to make it distributed. These components are a Communication Manager (CM), a Network Manager (NM), a Resource Manager (RM), and Global Interprocess Communication (GIPC). This paper addresses the communication mechanism in PODOS. In any distributed environment, communication appears to be the performance bottleneck. Thus, in PODOS, we have implemented a high-speed communication subsystem that short circuits the network protocol stack, and further performs packet multiplexing (Transmission-Groups) across multiple network interfaces thereby achieving a two-fold performance gain. In this paper we discuss the highperformance communication subsystem in PODOS and further analyze the performance gain achieved by comparing the variants of the PODOS protocol with traditional networking protocol. 1.0 Introduction A distributed operating system (DOS) is basically the cooperation among a group of machines interconnected by a network such that the group of machines appear to the user as a single operating system. With distributed operating systems, users are not aware where their files are stored; nor are they aware that their programs may be executed by remote machines. All resources within the network are managed in a global fashion using global mechanisms rather than local mechanisms [1]. A group of machines could cooperate for a variety of reasons. A few of them are (1) Resource Sharing, (2) Performance Enhancement, (3) Reliability, (4) Fault Tolerance and (5) Transparency [1]. Tens of distributed operating systems have been designed and implemented with various goals. Most distributed system designs are willing to compromise on performance. On the other hand, systems that are designed to be performance oriented make no attempt to provide a single system image. They provide a high-performance computing environment (e.g., Beowulf [2], Condor [3], etc). High-performance computing environments are designed to solve one class of problems, whereas a PODOS is designed as a general high-performance computing solution. With these issues in mind, we are designing a distributed operating system that is performance oriented (PODOS) [4]. As a secondary design goal, PODOS provides a resource sharing environment. PODOS is the interaction of two or more monolithic Linux [5] machines. The PODOS design has a number of key performance benefits. A few of these are:

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

The PODOS File System - Exploiting the High-Speed Communication Subsystem

Performance Oriented Distributed Operating System (PODOS) is a clustering environment, being built on a monolithic Linux kernel. PODOS augments very few components to the Linux kernel in order to make it distributed. These minimal components are the Communication Manager (CM), the Resource Manager (RM), the PODOS File System (PFS) and the Global IPC (GIPC). Each one of these components are desi...

متن کامل

PODOS -- The design and implementation of a performance oriented Linux cluster

PODOS is a performance oriented distributed operating system being developed to harness the performance capabilities of a cluster-computing environment. In order to address the growing demand for performance, we are designing a Distributed Operating System (DOS) that can utilize the computing potential of a number of systems. Earlier clustering approaches have traditionally stressed more on res...

متن کامل

Understanding the Requirements Imposed by Programming Model Middlware on a Common Communication Subsystem

In high-performance parallel computing, most programming model middleware libraries and runtime systems use a communication subsystem to abstract the lower level network layer. The functionality required of a communication subsystem depends largely on the particular programming model implemented by the middleware. In order to maximize performance, middleware libraries and runtime systems typica...

متن کامل

Measuring the impact of alternative parallel process architecture on communication subsystem performance

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 protoc...

متن کامل

Parallel transport subsystem implementation for high-performance communication

Requirements of emerging applications together with rapid changes in networking technology towards gigabit speeds require new adequate transport systems. Integrated designs of transport services, protocol architecture, and implementation platforms are required by forthcoming applications in high-speed network environments. The transport subsystem PATROCLOS (parallel transport subsystem for cell...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 1999