[H] from above

[Mardagg]

We are having a full out war with ARS, and having a good time doing it.

Looks like you guys are moving in on OCN pretty fast.. nice!

[phil]

Hehe, thanks man....we are coming for you very soon :rolleyes:

[Mardagg]

Originally posted by phil
Hehe, thanks man....we are coming for you very soon :rolleyes:

You'll have to pass ARS first. ;)

[zakelwe]

That Mardagg is a very perceptive guy :D

Good luck with OCN .. we'll all cheer when you sink their bath tub.

Regards

Andy@[H]

[Dave S]

hehehe....thanx dude....OCN their going DOWN

[siggy]

WE ROCK........We are going to sink that pirate ship and then go after the Italians.

WE ARE THE TEAM THAT ROCKS!!!!!!!!!!!!

[MechCD]

We are great!


I kinda feel sorry for arstechnica...... HardOCP just flew up, no warning, nothing. How the heck did they get that many users at once?

I have lost my anti OCN feeling a while ago, sorry to any pirates. I only had that attitude for 2 days i think.

Now for hard ocp..... well, its boiling inside me, they never talk to us, no taunts no "you guys will never catch us" . no anti [h] either jut a few gripes

[Mardagg]

Originally posted by MechCD
We are great!


I kinda feel sorry for arstechnica...... HardOCP just flew up, no warning, nothing. How the heck did they get that many users at once?

I have lost my anti OCN feeling a while ago, sorry to any pirates. I only had that attitude for 2 days i think.

Now for hard ocp..... well, its boiling inside me, they never talk to us, no taunts no "you guys will never catch us" . no anti [h] either jut a few gripes

ARS had plenty of warning, they are too busy running after ET tho.
and [H]ardOCP doesn't talk with you? huh, that's strange.. I could have swore I saw a couple of members from there posting in this message alone. ;)

[relic]

Come and get us TGC...You're a great team with an awesome attitude.

Nothing would make me happier than to be locked in a race with The Collective for #1 ( of course that means ARS would have to settle for third ;) )

Genome on, TGC. Hope to meet ya at the top!

-relic@[H]ardOCP

[MechCD]

there we go!

I feel better bout HardOPCP now :D

[phil]

Originally posted by relic
Come and get us TGC...You're a great team with an awesome attitude.

Nothing would make me happier than to be locked in a race with The Collective for #1 ( of course that means ARS would have to settle for third ;) )

Genome on, TGC. Hope to meet ya at the top!

-relic@[H]ardOCP



We are trying man....you just have so many members!! :D

[relic]

Originally posted by phil
We are trying man....you just have so many members!! :D

Didn't start out that way..... ;)

I'd bet you guys could up your production 25% just by resurrecting the old parts sitting in the closet!

So you guys gonna pass a few teams and settle for third?

[phil]

Originally posted by relic


So you guys gonna pass a few teams and settle for third?


Hell no, we ARE going to be number one.....the question is when?

We'll meet you there :D

[phil]

To be honest relic, when the next client comes out it will support http transfers. The majority of the Ars guys have access to vast amounts of corporate machines. You may have a bigger fight on your hands than you think ;)

[siggy]

Wow that will be something to keep an eye on. It could change a lot of standings. We don't know who is behind us that may have a few Corp. machines laying in waiting.:eek:

[phil]

Originally posted by siggy
Wow that will be something to keep an eye on. It could change a lot of standings. We don't know who is behind us that may have a few Corp. machines laying in waiting.:eek:


We have a few members on this team that it could be advantageous to have http support :D

[MechCD]

I'm out of parts in the closet :( at leats any that will do a gene in a week :)

How bout that there 90mhz pentium 1?

[Bruce]

Originally posted by MechCD
I'm out of parts in the closet :( at leats any that will do a gene in a week :)

How bout that there 90mhz pentium 1?

Me too. I can't afford the electricity to run the 486's in my closet.

[relic]

Originally posted by phil
To be honest relic, when the next client comes out it will support http transfers. The majority of the Ars guys have access to vast amounts of corporate machines. You may have a bigger fight on your hands than you think ;)

Been hearing that for a long time. ;)

They claimed the same thing would happen when the client was stable. .99 is here and stable, you see any difference? ;)

Talk is cheap, it's the results that find the cure.

Genome on Phil and TGC!

[Dustin]

Originally posted by relic
Talk is cheap, it's the results that find the cure.


I agree, but I don't think it's all lies. Most of the talk is just wishful thinking, and not even knowing if they can run these machines before bragging all over the place.

I know I'm guilty of hollering about running 15 dual PIII’s, which in the end didn't pan out. :(

[relic]

Originally posted by Dustin
I agree, but I don't think it's all lies.

Yeah, I don't think it's deliberate "lies" either, just wishful thinking and bravado.

Still don't generate WU's though ;)

Gnome On!

[zakelwe]

If you want insults :-

" My mother in law genomes more than you do in a week. And she doesn't even have a computer "

:D

ARS might get a big influx when TBR or whoever finds the RC5 key first ( I'll give them a clue ..it's down the back of the sofa) and they go over to genoming

If they sorted themselves they would anyhow. Then it will be neck and neck again.

The poor old Italians have gone from 1st to looking at 5th ! Must be pretty disheartening...

Regards

Andy@[H]

[Virus]

http support will most certainly go in Ars's favor. Mark, my words when I say this, "The Knights Who Say Ni! will be a team to deal with when http support arrives". Many of their seti folks are waiting for this to happen and they, like Ars, have lots of corporate pc's at their disposal. A couple have just recently jumped the seti ship and their output is putting them around 7th place at the moment. http support will definately conger more competition and help this and the folding project, but will hurt the current teams that are made up of people with home pc's, which is TGC for the most part. I for one am very happy here and don't care if we slip to 20th when http arrives. Hopefully it will attract more home computer geeks to join us!

[Rizzo]

Sorry to seem stupid, but what will HTTP support allow that makes such a big difference? In other words... what is it?

[phil]

Originally posted by Rizzo
Sorry to seem stupid, but what will HTTP support allow that makes such a big difference? In other words... what is it?


It allows traffic to be routed on port 80 (www). This will allow larger companies who are behind firewalls to access the internet to send and receive work units where they normally couldn't get past the firewall.

[Dave S]

Originally posted by relic

So you guys gonna pass a few teams and settle for third?
hehehe third!!! no way we want that #1, give us some time & we’ll see wot we can do........:D
cheers man CYA @thetop :cool:

[Mardagg]

You guys have the right attitude, that's for sure. What's so great about this, is in the end, it's good for all of us.

[Dave S]

Mardagg, respect man, yer by doing this as you say it helps everybody........& we all have a LOT of fun & make some friends, hey this is :cool:

CYA @thetop:D

[Jodie]

Here's a question. Why don't we simply write a 'middleman' proxy?

Here's the thought: A "server" that accepts the connection, running concurrently on the machine with the genome process - takes port 10100 or whatever as input, it then opens an http connection to the outside world, and let's say I put an http server on the public net - I have two full T-1s into the house, I'd be willing to do that for the team. - and it receives the port 80 connection. It takes the resulting data, severs the connection, strips the http header off of it, reopens a connection to stanford just like a normal genome connection, and flings the data at it, posing as the user that orginally sent the WU. Not a good description - I need a whiteboard. Anyway, I could whip that together for Linux atleast in a couple hours. Any windows programmers here? I could write it in ANSI C, you'd just need to get a winsock->berkeley sockets layer going. Do we have people here sitting on a reasonable amount of corporate processing power and waiting for that functionality? It's a hack, but it sure as heck would work...

[Bruce]

Originally posted by Jodie
Here's a question. Why don't we simply write a 'middleman' proxy?
. . . . It's a hack, but it sure as heck would work...

No question it would work. The only real question is: Is it worth the effort. Stefan will get it running in the native client soon (the tools are now available -- built in to the interface product they bought), so the only questions are how long would it take to get it running and would you be happy to discard it "soon"

Another useful product would be a simplequeueing proxy server -- like SetiQueue and other similar seti add-ons -- but even that is probably unnecessary work. If vspxx.stanford.edu goes down, you're not dead in the water the way you were with seti when berkeley went down.

If you wrote either one (or combined them) you'd be locked into supporting it and adding nice features like keeping logs of who got what protein, who returned what gene, setting queue priorities for big or small proteins, etc., etc. etc.

I think it is a great idea -- just be sure you know how big a project you're willing to tackle.

[Jodie]

Actually, my goal would be a quick hack to 'tide over' until Stephen and company get the native support done. Just a few hundred lines of code, most of it from projects I've done in the past, strung together. The only way it would make sense to me is if someone or groups of someones comes back from the team and says they have 20+ Ghz waiting for it. If they do, then it makes perfect sense to give the team the month or more lead before native proxy-ing is released... If I were going to write a full fledged too (and I'm actually working on one right now for this) it would be a significant control utility for hetrogenous networks of machines like what I have at home. Basically monitoring the processes on each machine, reporting status wirelessly to my palmtop and allowing me to control them from the palmtop anywhere I have CDPD or 802.11 access. It will also page my cell phone and in the immediate future, allow me WAP access to control them... All without any kind of GUI on the cruncher, interfaced to native HTTP/HTML1.0, to keep the load bearing on the system down. THAT'S my big project...

[wbierman]

I have 350GHz to 425GHz (that does not include servers - 15 dualies) just waiting to access port 80. Your idea sounds terrific BUT, I have no control over the Firewall. It is controlled by another group which means I have to wait for the client to natively support it.... 325 of these machines are brand new PIIIs @ 933MHz running Win2K Pro.

[Bruce]

Yep. That is a big -- and valuable -- project.

Regarding the socks proxy, Isn't that the same thing they're talking about here using NEC's Sockscap and HTTPort3?
http://www.hardforum.com/showthread.php?&threadid=119344&perpage=15&pagenumber=3

[Jodie]

Originally posted by wbierman
I have 350GHz to 425GHz (that does not include servers - 15 dualies) just waiting to access port 80. Your idea sounds terrific BUT, I have no control over the Firewall. It is controlled by another group which means I have to wait for the client to natively support it.... 325 of these machines are brand new PIIIs @ 933MHz running Win2K Pro.

Ok, THAT's worth it...

I don't think you will have problems with the firewall. Let me see if I can explain my architecture idea a bit better...

You actually execute TWO things on each machine. (or the little servelet on a machine inside your firewall that everything can access if that is allowed in the security policy) You run g@h and tell it that your server is 'localhost' and the port is whatever. The second thing is a tiny servelet, much lighter than a proxy server. The tiny servelet running on localhost will take the connection from the client, take the results, close the connection, package them up in http packets, and ship them out over port 80 source/dest to a machine sitting in the outside world, say for example on my home T-1's. It will hand the work off to a server on that outside world and and that server will hand it a cached WU to deliver to the client as if it were a standard Stanford connection, then drop its virtual http "connection". The master server and client-servelet will then strip out the respective http headers. The server will make a valid connection to stanford identifying itself as the user that gave it the packets and pass them off. The servelet will hand them back to the client as if they came across a 10100 connection from Stanford. I just did a capture on a session with stanford - the negotiation would be trivial to emulate.

Does that make more sense?

[Jodie]

Originally posted by Bruce
Yep. That is a big -- and valuable -- project.

Regarding the socks proxy, Isn't that the same thing they're talking about here using NEC's Sockscap and HTTPort3?
http://www.hardforum.com/showthread.php?&threadid=119344&perpage=15&pagenumber=3

Yup! Leave it to me to reinvent the wheel all the time... Sigh...:rolleyes:

Seriously though, that method looks a wee bit heavy. If you're trying to be at all stealthy, I wouldn't recommend it. I was going to go for the hyper-lite console type method with file logging. Something that hideit could get rid of nicely. But, hey, their tools exist today! If there's a reason that won't work for Wbierman, then I'll crank mine out tonight for him to play with and see if it will... 325Ghz is 325Ghz. Of course, then I'll have a REALLY icky moving target to shoot at...:rolleyes: :D

[wbierman]

Believe me when I say this... I'm in no hurry to run out and load the client on hundreds of machines.

What I really need is a monitoring app that can sniff the network and report on machine status with regard to "is the client running?" My pipe out is a T3! But I still have to stagger clients so I don't have massive dumps.

Your idea is very clear now and would give us a huge head start against the others....

[viperdog]

Originally posted by wbierman
I have 350GHz to 425GHz (that does not include servers - 15 dualies) just waiting to access port 80. Your idea sounds terrific BUT, I have no control over the Firewall. It is controlled by another group which means I have to wait for the client to natively support it.... 325 of these machines are brand new PIIIs @ 933MHz running Win2K Pro.

You are running the client on those in -nonet now right? I realize you'd have to do a lot of copying to set them up but hey when the new client comes out or Jodie puts up her solution then WOW..what a dump that would be...

[Jodie]

Originally posted by wbierman
Believe me when I say this... I'm in no hurry to run out and load the client on hundreds of machines.

What I really need is a monitoring app that can sniff the network and report on machine status with regard to "is the client running?" My pipe out is a T3! But I still have to stagger clients so I don't have massive dumps.

Your idea is very clear now and would give us a huge head start against the others....

I can understand not wanting to run out and install it on hundreds of machine but all MIPS (Meaningless Indecies of Performing Speed) count, right? Think I'll do it.

The monitoring app you're describing is exactly what I've been working on, taken one level deeper: Allowing you to do exactly that with a WAP enabled cell phone or wireless enabled palmtop... I'm testing the first phases now. I haven't decided how it's going to deal with Windows clients yet - but I know it will..

[wbierman]

No...no -nonet occuring at this time. All are in the process of having their images replaced. Or the machine is being replaced. I have limited acces during the summer. (I don't have all the keys) I do have access to servers and have been upgrading the network all summer. The old pipe out was a fractional T1...now it's a full T3. Moved from 13 seperate Netbios domains, a Linux Firewall- DHCP, with a Netware 3.1 Directory Server - to Win2K with Active Directory spanning 4 physical "sites" with 3 logical colission domains within each site. Still have the Linux Firewall which was replaced a couple of months ago because it would crash at least once a day with an average down time of 12hrs per pop.

I have not had "time" to fuss with -noneting. Running clients would require significant monitoring and maintainance. Finally we have client .99 that seems to be stable.

Once I start this, then all will expect results...huge results = huge responsibility...

[Jodie]

Yup! I would be hesitant about undertaking that... The trick is to just install a dozen or so and not tell anyone where it's coming from or that there's anything else available...;)

[wbierman]

Actually installation is the easy part. Just create an install package and deliver it within a group policy. Click Apply and out it goes, replicating to the DCs and then out to the clients which then reboot as part of the package setting it as a service in Win2K and starting the genome client.

Its the monitoring part and the maintainance....

[Jodie]

Nod. Not refering so much to the deployment as to maintainence. Although you should be able to deploy a scheduled task to maintain it, right?

[wbierman]

Indeed.

[pelligrini]

Ya'll might talk to dyyryath. I believe he has already done a lot of work on an app very similar.

[relic]

Genome Fleet Manager is Dyyryath's application for deployment and management on NT/2k networks.

He's a good guy and I'm certain he'd send you a copy. He frequents the mirky depths of the ARS DC forum. You could find him there. ;)

[Mardagg]

Looks like ARS has answered the call, this will be a wonderful battle!

[Dave S]

Yep I noticed that too, good for the project keep it up, well done to all involved, you rock!!!:cool: