Pavana Yalamanchili (pavanay@hotmail.com)
Thu, 04 May 2000 07:55:23 PDT

The Sprite Network Operating System

Networks, large memories and multiprocessors are the three general trends in
computer technology, which motivate the building of a new operating system,
in this paper, Sprite. In Sprite, the goals were to hide the distributed
nature of the systems as much as possible, while providing the sharing and
communication of time-shared machines, to manage physical memory in a way
that maximizes the potential for file caching and to facilitate the
development of multiprocessor applications, and that the operating system
will itself be able to take advantage of multiple processors in providing
system services. Thus the overall goal for Sprite was to provide simple,
efficient mechanisms that capitalize on these three technology factors
affecting the systemí design. Sprite implements a set of kernel calls that
provide sharing, flexibility, and high performance to networked
Some of the interesting features of the kernel implementation are:
∑ The kernel contains a remote procedure call facility that allows a
workstationís kernel to invoke operations on other workstations, to
implement features such as network file system and process migration.
∑ Sprite uses a simple mechanism called prefix tables to manage the name
∑ Sprite caches file data on both server and client machines
∑ The virtual memory system uses ordinary files for backing storage,
simplifying implementation, facilitating process migration etc.
∑ Sprite guarantees that processes behave the same whether migrated or not.
I think this was a well-written paper and all features discussed at the time
were operational (except for the one feature related to migration which was
still under development). Process migration in Sprite is transparent to both
the user and the migrated process. Finally, the authors summarize how they
hope that Sprite will provide sharing, flexibility and performance to

The Distributed V Kernel And Its Performance For Diskless Workstations

The goal of this paper is to describe the performance evaluation of the
distributed V kernel for diskless workstations. Distributed V kernel is a
message-oriented kernel that provides uniform local and network interprocess
communication. The authors conclude that it is possible to build a
distributed system with all network communication using the V message
facility even if the network nodes have no secondary storage. Their
measurements of the V kernel were primarily concerned with the cost of
remote operations versus the cost of the corresponding local operations and
the cost of file access using V kernel remote operations versus the cost for
other means of network file access. The performance study showed that V
kernel IPC performed satisfactorily despite its generality. I feel the
authors did a good job of justifying their work and pointing out the
tradeoffs but I think the authors pursued only certain areas of interest. As
is said in the paper, V kernel does not provide all the functionality of the
LOCUS kernel but from the specific comparisons with LOCUS, the authors
expect its overall file access performance to be comparable to LOCUS running
on the same machines and network. At the time this paper was written, V
kernel was in active use at Stanford and at other research and commercial
establishments. More experience and measurement of file server load and
workstation file access is required to decide if file server processing is a
significant problem in using diskless workstations. Finally, the paper
compares the V kernel performance to that presented by other authors, to
Accent and to LOCUS.

Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com