|
Abstract - Technical Program - ES 99
Massively Distributed Systems: Design Issues and Challenges
Dan Nessett, 3Com Corporation
Abstract
The forty year trend in the
computing industry is away from centralized, high unit cost, low unit
volume products toward distributed, low unit cost, high unit volume
products. The next step in this process is the emergence of massively
distributed systems. These systems will penetrate even more deeply into the
fabric of society and become the information power grids of the 21st
century. They will be ubiquitous. Most will operate outside the normal
cognizance of the people they serve and most will be based on embedded
systems that present non-traditional computing interfaces to their users.
They will be engineered to operate as distributed utilities, much like the
energy, water, transportation and media broadcast businesses do today. The
first deployment of massively distributed systems is likely to occur as
support structures for these industries.
Massively distributed systems will differ from existing
distributed systems in important ways. Such systems eventually will
interconnect billions of nodes. This will necessitate changes in the way
nodes interact with one another. One-to-many communications will be the
norm, rather than one-to-one. The size of on-line application communities
will necessitate the use of statistically correct rather than deterministic
algorithms for resource accounting, fault detection and correction, and
system management. These communities will coalesce and dissolve rapidly in
order to host events that are of interest to groups formed specifically for
this purpose. This will require new approaches to naming, routing, security
and privacy, resource management and synchronization. Heterogeneity will be
even more of a factor in the design, implementation and operation of
massively distributed systems.
This paper explores the nature and characteristics of
massively distributed systems by proposing some examples and then using
them to characterize nine major design issues. Seven of these are drawn
from seminal work in the area of distributed systems. Two others are based
on experience in distributed system design and implementation subsequent to
that work.
- View the full text of this paper in
HTML form and
PDF form.
- If you need the latest Adobe Acrobat Reader, you can download it from Adobe's site.
- To become a USENIX Member, please see our Membership Information.
|
Need help? Use our Contacts page.
Technical Program