From: Tarik Nesh-Nash (tarikn_at_microsoft.com)
Date: Mon Feb 02 2004 - 18:14:53 PST
This paper shares the experience of the implementation and the use of a distributed system: Grapevine. It is interesting to analyze the different encountered problems and understand how these problems were addressed since then.
The paper starts by presenting an overview of the architecture of the system and its substantial growth. These will be very important to understand the different problems that are addressed later.
The paper discusses the decision making process regarding scalability, performance, transparency, load balancing and reliability. Most importantly, it presents the learned lessons from using this system in the production field for few years. Long term planning and good design are important; bad design may be setting limitation to the users growth and the network growth. Also, legacy seems to be an issue as the system grew and requirements changed.
This document is a great resource for new distributed system designers. I would expect an important number of responses and articles to try to find better solutions to the different problems presented on this paper.
This archive was generated by hypermail 2.1.6 : Mon Feb 02 2004 - 18:15:14 PST