[rc5] [ADMIN] futures past... distributed.net

Adam L. Beberg beberg at distributed.net
Thu Oct 23 11:59:07 EDT 1997


Greetings, and congratulations to everyone, root beer for all...

Its time once again for me to emerge from behind my monitor and shed some 
light on what's happening with distributed.net, v3, and all that jazz...

First I would like to congratulate everyone that ran RC5-56 clients for 
the last many months. This is a HUGE shared victory for all of us, proving 
that 56bit encryption is too weak, and that distributed processing on the 
Internet is here to stay... I can't thank the people involved enough, 
Bovine for all his coding, Nugget for keeping the lists and press under 
control, and all the coders which I won't even try to name for fear of 
leaving someone out...

Though I never expected to say this back when my stats (pre Nugget SQL) 
said we would need another 138 years to solve RC5-56, we solved it a
little too fast, and it caught us slightly unprepared. Although we have
started rolling out RC5-64 clients with most of the features that 2.005
had, this is just an interim step until we get the v3 network in place. 

First let me say that v3 will be a big step away from the way 
distributed.net has operated in the past. From the start I'm designing v3 
as an open protocol. This means that client development, and the writing 
of cores, will be open to anyone with the required skill and motivation. 
v3 is a full protocol specification for distributed computing, not just a 
new client. You will also eventually be able to run your own server/proxy 
for academic projects, so no more waiting 3 weeks for your thesis project 
to process :)

While I'm documenting v3 as quickly as I can, I do want to be sure 
everything is in order before I open it up for peer review. Once it's 
done, people will see it. Once everyone is happy with it, it will be 
carved into stone and coding will begin. When I designed the v2 clients, I 
was crippled by backward compatibility with the v1 clients. I consider 
this a fate worse then death (just ask Intel). v3 is a fresh start and I'm  
not going to rush development.

At this point (meaning after all this excitement dies down) I will be 
disappearing again to go finish off the protocol and client/server 
specifications for v3. Then while I code the v3 servers and support 
layers, the distributed.net coding team (and anyone else who's interested) 
will begin coding clients and modules [cores]. Hopefully, about the same 
time I wrap up the servers we will begin to see client modules ready to 
commence processing. Obviously, 64-bit RC5 will be the first of the list 
of cores, but it is not the only project we have planned. We're also 
linking it all in to the powerful stats engine that Nugget will be 
developing *grin* and distributed.net will be the lean, mean (in a nice 
way), and powerful entity that we've all envisioned.

v1 was a proof of concept, v2 tested out our proxy system, and v3 will be 
the tool we use to change computing as we know it...

I've lately been in contact with the organizers of the Mersenne Primes
search, and they are very enthusiastic about having us working on Mersenne
and we will begin collaborating on development of the necessary V3 cores
for a distributed.net Mersenne Primes effort. I'd like to thank them for
sharing my vision for distributed.net, and realizing that we're all in
this together. Their enthusiasm and cooperation is truly refreshing and
our combined efforts are turning out to be a smooth and seamless process. 

Mersenne primes as you know will be the first real project of v3. Mersenne 
hunting requires quite a lot more horsepower then RC5 does tho, and so 
only our faster machines will be working on it, the slower ones will 
continue with RC5-64, or whatever project you select. The clients will be 
fully configurable to select either one project, or choose those best for 
it's capabilities. Alpha's for instance will excel at Mersenne due to the 
large amounts of floating point computation. As we all know Alpha's don't 
do RC5 very well, so v3 will allow each client to get the most out of it's 
relative computational strengths.

I ask you for your patience. The distributed.net team is working to 
provide v2 RC5-64 clients as quickly as possible, along with proxies, 
stats and all the rest. Download them now, run them, use the victory to 
recruit more people. Each and every one of you is a critical part of 
distributed.net, and we couldn't do anything without all your idle cycles.    

We've proven we're here to stay, and much greater things are around the 
corner. If v3 is everything I'm designing it to be, someday you'll be 
telling your kids you where there when _thee_ Net began. Prepare to be   
amazed, and welcome to the future :)

- Adam L. Beberg
  beberg at distributed.net
  distributed.net coordinator
  http://www.iit.edu/~beberg/

----
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.



More information about the rc5 mailing list