The Artima Developer Community
Sponsored Link

News & Ideas Forum (Closed for new topic posts)
Distributed Computing Economics

2 replies on 1 page. Most recent reply: Jul 11, 2003 1:17 PM by Talip Ozturk

Welcome Guest
  Sign In

Go back to the topic listing  Back to Topic List Click to reply to this topic  Reply to this Topic Click to search messages in this forum  Search Forum Click for a threaded view of the topic  Threaded View   
Previous Topic   Next Topic
Flat View: This topic has 2 replies on 1 page
Frank Sommers

Posts: 2642
Nickname: fsommers
Registered: Jan, 2002

Distributed Computing Economics Posted: Jul 2, 2003 2:13 PM
Reply to this message Reply
Advertisement
Computing economics are changing. Today there is rough price parity between (1) one database access, (2) ten bytes of network traffic, (3) 100,000 instructions, (4) 10 bytes of disk storage, and (5) a megabyte of disk bandwidth. This has implications for how one structures Internet-scale distributed computing: one puts computing as close to the data as possible in order to avoid expensive network traffic.

http://www.clustercomputing.org


Dan Creswell

Posts: 49
Nickname: dancres
Registered: Apr, 2003

Re: Distributed Computing Economics Posted: Jul 8, 2003 4:55 PM
Reply to this message Reply
Good article, IMHO. Clearly articulates where Grid computing is likely to provide benefits and why.

The downloadable code aspects are certainly interesting from a JINI perspective but I think there's a slightly different philosophy at work.

JINI has always had the concept of downloadable proxies and with good reason as it allows a client to start using a service given just an interface description.

It's often been suggested that such a proxy might actually contain the service itself but, in a lot of cases, the proxy communicates on the client's behalf with some back end (potentially stateful) service.

So, the client downloads the proxy once when it starts to use the service (and may download it again in the future but not very often IMHO) and does only a little computation. Most computation goes on in the backend which is, typically, close to it's data/state/storage.

Now, thanks to multicast discovery, it's easy to have the client only locate services in the near network vicinity (just use a small TTL) rather than try and use those on the other side of the world, in network terms. Thus the client potentially has only a few hops to the service which should be close to it's data.

In conclusion, I think JINI already provides some useful mechanisms to get a handle on these economics it's more an exercise in whether we use them or not.

It'd be interesting to take this stuff and consider the implications for wide-area clustering/replication issues....

Talip Ozturk

Posts: 103
Nickname: talip
Registered: May, 2003

Re: Distributed Computing Economics Posted: Jul 11, 2003 1:17 PM
Reply to this message Reply
> It'd be interesting to take this stuff and consider the
> implications for wide-area clustering/replication
> issues....

i have written couple of entries on clustering/replication issues which is available at <a href="http://freeroller.net/page/talipozturk">http://freeroller.net/page/talipozturk</a>.

Flat View: This topic has 2 replies on 1 page
Topic: Can't get there from here Previous Topic   Next Topic Topic: The Zen of Python


Sponsored Links



Google
  Web Artima.com   

Copyright © 1996-2014 Artima, Inc. All Rights Reserved. - Privacy Policy - Terms of Use - Advertise with Us