Convergence of P2P and Grid Predicted 117
tom_conte writes "From the proceedings of the 2nd International Workshop on Peer-to-Peer Systems (IPTPS'03), "On Death, Taxes, and the Convergence of Peer-to-Peer and Grid Computing" compares the two current popular incarnations of distributed computing technology, Peer-to-Peer (P2P) and Grid Computing. It also predicts the convergence of the two technologies: "The complementary nature of the strengths and weaknesses of the two approaches suggests that the interests of the two communities are likely to grow closer over time." This paper is worth reading if you want to clear up the marketing cloud that surrounds these two technologies and sometimes makes them hard to distinguish."
Sounds like... (Score:5, Insightful)
Grid computing can survive just as well without P2P. I'm not so sure that it's the same in reverse.
Re:Sounds like... (Score:2, Informative)
Seti@home (Score:3, Insightful)
It is a serious well supported argument. You are just shooting your mouth off with ad hominiem attacks which probably aren't valid.
Actually (Score:3, Insightful)
If we redefine P2P from being a way of copying software and music to a way of sharing computational code across a network, then it all becomes so much more acceptable.
It's a conference for P2P. Did you really think they'd come out and say that it's a hopeless dead end? Did you expect they'd say that unless they can justify it's existence that P2P will be called a piracy tool?
Re:Seti@home (Score:3, Informative)
Seti is plain client server.
My seti client doesn't talk to your seti client. They both report back to a central server.
Re:Seti@home (Score:1)
Re:Seti@home (Score:2)
Does that mean a webserver is peer to peer?
No.
A fileserver is quite useles without any clients to store files on it and read files from it.
Does that mean a file server is peer to peer?
No.
Peer to peer means you have peers communicating with peers.
In seti you have a massive specialised central server and you have clients. The clients aren't at all a peer of the server. It's client/server computing not p2p computing.
Seti would only belong in the "P2P realm" if it was P2P and it's not!
Re:Sounds like... (Score:5, Insightful)
Really -- if you want examples of P2P use that aren't the often over-hyped file-sharing scenarios, imagine an FTP server that can scale really well, i.e. the more people who are using it, the more capacity it has. Imagine SOMA FM being able to stream to orders of magnitude more people than it does now, by having clients on fat pipes reflect the stream they're getting to others. Sure, there are still problems to solve in this case (latency, especially for live-as-it-happens type content) but the potential, I think, is incredible.
Regards,
John
Re:Sounds like... (Score:3, Informative)
P2P is a solution to massive content distribution; any kind of content. And there are different scales. P2P may show itself in many ways in the near future, whether you're chillin' to SomaFM [somafm.com], backing up your data to/from a redudant multipeer network, or maybe distributing large content from a web site and benefiting from clients' bandwidth for peer to peer redistribution.
my humble opinion (Score:1)
The future of the Grid (Score:5, Interesting)
In the aftermath of the dot com crash, companies are falling over themselves trying to snag onto the "next big thing".
Now we have two different worlds colliding, with people pushing 'em that have been ignoring each other all this time.
They've at least recognized this, however, there's still a HUGE problem.
They can't make it easy for the average person to install and use.
They (the Grid folks in particular) seem to be missing this, big time. Globus is NOT easy to install...it's not an out of box experience like any of the P2P things are. It's a multi-day install, and you have to know what the heck you're doing.
Secondly, the world doesn't need yet another Corba-like thing to make everything interoperate with everything else with MORE glue on top of it. KQML should have taught people this lesson back when that was all the rage in agent systems. If you want two systems to talk to each other, couple 'em in whatever language you want and stick to it.
There's so much extra overhead in doing tasks that "the grid" is supposed to take care of....man, I wish these people would just sit back and take notice of the other distributed systems out there that are out there and working and solving problems without foisting yet another distributed computing paradigm (oh hell, I can't believe I used that word...forgive me), on the world.
Lord knows we don't need it entangling reasonably well put together P2P systems with the tentacles of the heavy-weight "Grid".
Re:The future of the Grid (Score:5, Interesting)
Apple seems to be on top of this with technologies such as TCP/IP over Firewire and most notably, Rendezvous technology. The potential here is amazing.
Re:The future of the Grid (Score:3, Informative)
Ah, the type of post I expect from an AC. However, for the benefit of others with at least half a brain here, check out this link [apple.com] to find out all about TCP/IP over Firewire. It's pretty impressive and provides a fast easy way to move lots of data in an easy to implement network combined with auto detection and configuration.
Re:The future of the Grid (Score:4, Informative)
I should have remembered this for my immediately prior post, but actually, I wrote an article about the implications of distributed computing including a discussion about TCP/IP over Firewire back in December. You can read about it here [applelust.com]
What's TPC? (Score:2, Funny)
Re:The future of the Grid (Score:5, Informative)
The grid has been trying to gear up from academia/research for so long, I can't even remember when I first heard of it.
About eight years.
In the aftermath of the dot com crash, companies are falling over themselves trying to snag onto the "next big thing".
Please enumerate these companies. I'd love to consider them as possible customers. Currently grid computing is mostly met with, well deserved, skeptisism.
Now we have two different worlds colliding, with people pushing 'em that have been ignoring each other all this time.
Grossly unfair. At least from the grid side we've long considered p2p and there are even working groups discussing this intersection in GGF.
They can't make it easy for the average person to install and use.
Bullocks on two counts.
1) *average people* don't and shouldn't use gridware yet.
2) The dominent gridware is trivial to install compared to the problems it solves. It's like saying that you want to take a crack a proving an ancient theorem but you don't want to have to learn any math. TS.
They (the Grid folks in particular) seem to be missing this, big time. Globus is NOT easy to install.
Again, wrong on two counts.
1) Globus is not the grid. Globus is *a* project developing gridware.
2) Globus is fucking trivial to install (and I'm not on their staff). Maybe not as easy to install as kazaa, but then gridware is targetted at people who want to do more with the internet than look at pictures of Gillian Anderson's pussy. That said, review point one. There are other grids available.
Secondly, the world doesn't need yet another Corba-like thing to make everything interoperate with everything else with MORE glue on top of it.
Now there's an almost complete lack of content. Not a single grid development house is going in this direction. There is a *broad* consensus that using open communications protocols (i.e. web services) is the way to go instead of making something proprietary.
KQML should have taught people this lesson back when that was all the rage in agent systems.
Right, because agent systems and grid are *so* similar.
Come on now, KQML was a language for doing REST (to a first approximation) a grid is a far more complex concept.
If you want two systems to talk to each other, couple 'em in whatever language you want and stick to it.
Again, a complete and utter lack of understanding what the grid is about. Your comment amounts to having a cabal of grid programmers bless a particular language and then demand that everyone write to it. Dumb. Not that there aren't language bigots in the grid community that would do that if they could, but dumb nonetheless.
man, I wish these people would just sit back and take notice of the other distributed systems out there that are out there and working and solving problems without foisting yet another distributed computing paradigm (oh hell, I can't believe I used that word...forgive me), on the world.
I wish luddites would do a bit of reading and educate themselves before assuming that everything they didn't come up with is nonesense. Especially luddites who have *no* idea the depth of the library at my company.
The grid solves problems that exist and aren't being solved in other ways except through enormous investment by each and every company that wants to solve them for themselves.
Another analogy. Your comment is akin to demanding that instead of adopting Windows (which is a hassle to install, run and keep secure) that they instead write their own operating system tuned to their own needs. Sillyness.
Hugs and kisses from the future.
Gridware (Score:2)
In the practicle department, we have two technologies we've used for grid computing. I'm going to guess you work on one we've used, Sun Gridware (now opensourced?) used to be Codine. It worked well enough. The interface was pretty easy, but for the users, kicking off programs to run randomly on a grid of machines was not as easy as...
Mosix (or OpenMosix). Now I'm sure theres a hundred good reasons why it doesn't qualify as gridware, but that is how we use it. It was simple to install and monitor also. Hidden enough, the engineers don't even know they are using it. Thats the kind of plug and play every IT manager wants.
The convergence, I agree is going to happen. But honestly today people just haven't become sophisticated enough to expect it yet. Or even want to. Perhaps it won't happen in our lifetime.
-------------
OnRoad [onlawn.net]: Boldly searching for the efficiency our engines deserve.
Re:The future of the Grid (Score:4, Funny)
If thats not a marketing slogan, I don't know what is.
Re:The future of the Grid (Score:2)
I'd say I know who you are, but you use both more profanity and punctuation than the person I'm thinking of. So...hi.
Re:The future of the Grid (Score:3, Interesting)
I mean companies like IBM who are looking for the same customers you are.
Globus is fucking trivial to install
Unless Globus has gotten a helluva lot easier to install in a recent release, "fucking trivial" is a load of bullshit. I've done, and other people I know have done, several installs each, with people that knew what they were doing, and I'm telling you, even with that help, it took at least two days of farking around to get it to install properly.
So I call major bullshit on "fucking trivial"
Right, because agent systems and grid are *so* similar.
Come on now, KQML was a language for doing REST (to a first approximation) a grid is a far more complex concept
No shit. The panacea of KQML was that "everything will talk to everything else", just like the grid, and Corba. You know what? Everyone dropped KQML because they realized they DIDN'T need to talk to everyone else. They just needed to talk to themselves. You end up with so much extra baggage because of this it ends up being a complete pain in the ass.
Again, a complete and utter lack of understanding what the grid is about. Your comment amounts to having a cabal of grid programmers bless a particular language and then demand that everyone write to it. Dumb. Not that there aren't language bigots in the grid community that would do that if they could, but dumb nonetheless This completely misses the point. The point isn't to have everyone write in the same language, it's to have a simple system for people to use. The extra hurdles that people have to get their programming language usable on the grid aren't trivial, and not every language is supported. It adds complexity to the system that shouldn't be required.
The grid solves problems that exist and aren't being solved in other ways except through enormous investment by each and every company that wants to solve them for themselves. Well, now's the time for me to call bullshit. No one is doing this now...they might in the future, but I seriously doubt it. Companies aren't going to be forking over their computations to other companies to do, when they can do what they've been doing in house. More hardware might be sold, but it won't be shared.
In the end, if the grid does actually get used anywhere, it'll be in-house, or in academia. And even there, people are arguing which place is the right place to host most of the activities...which, is completely STUPID, IMHO, because it completely misses the point! I've been in those meetings where they're deciding "where we'll concentrate our X computations" and "site Z will be where we concentrate our Y capabilities". Holy living crap! I've never seen anything so farked up in my life. There are a lot of really pissed off people here because of that.
Another analogy. Your comment is akin to demanding that instead of adopting Windows (which is a hassle to install, run and keep secure) that they instead write their own operating system tuned to their own needs. Sillyness
This completely makes my point. We don't need another one of these things. There are plenty of lighter weight alternatives out there which people have developed and have been using for the last few years. One great example is Cisco's Spanish Inquistion built on top of Jini. (spare me the language bigotry...not you...the other readers of this).
I wish luddites would do a bit of reading and educate themselves before assuming that everything they didn't come up with is nonesense.
And double at ya....most of the last eight years (and my involvement has been for the last six) has been pedaling this crap. The dot.com crash is what made people latch onto this. Before that, you couldn't give it away...which is ironic, because they were.
TV and Computers will converge (Score:2)
I used to Believe it...(not any more)
but you know what this convergence of P2P and Grid sounds logical...I choose to believe!
Re:TV and Computers will converge (Score:1)
Consider the possibility that some things take longer to converge than one might expect.
Re:TV and Computers will converge (Score:2, Insightful)
My ISP and my cable company are the same legal person, same bill--and I can even rent a specalized "TV computer" from them if I want to.
Wait until digital HDTV becomes prominent, and network-wired houses are as common as telephone lines today. TV and PC will converge--it's just going to move along a multistep process at the speed of the slowest partner.
Killer app of the '00s (Score:3, Interesting)
Ecch.
Re:Killer app of the '00s (Score:3, Informative)
Re:Killer app of the '00s (Score:2)
I was at a user meeting for a 'grid computing' company, where they had a bunch of customers come in and give presentations about how they were using it in their environment.
Every single person that got up had a different definition for grid computing. Even employees of said company had different definitions.
This is exactly like saying something is 'web enabled' back in 1995. Everything had to be web enabled, and everything was, so long as your definition was very vague.
Re:Killer app of the '00s (Score:2)
Re:Killer app of the '00s (Score:1)
When the volume of the app outweighs the www, you know something BIG is goin' on.
The real issue is the desire for decentralisation. That desire is irrefutable.
Grid Computing, (Score:5, Insightful)
It seems as though this system would inherently be P2P. It's good to know the P2P people are starting to realize that there is more the P2P than file sharing. As for the grid people, they knew their system could be called "Peer to Peer" all along.
Re:Grid Computing, (Score:2, Funny)
Sounds like Borg computing.
Re:Grid Computing, (Score:1)
Re:Grid Computing, (Score:5, Funny)
Anybody else read that and get a vision of the Borg? Are we headed that way?
Heh. Just found that interesting. Lots of stuff talking to lots of other stuff. Emergent properties are bound to appear, right?
Re:Grid Computing, (Score:5, Insightful)
Then again, perhaps it is a case of "If you build it, they will come."
Re:Grid Computing, (Score:2)
Re:Grid Computing, (Score:5, Informative)
What it says, as far as I see it, in a nutshell:
In that sense, it seems like a really great paper, the kind of thing I've sought after for a while now. Nice one, Ian!
we already have that (Score:3, Informative)
The difficulty with this is not doing it, it's doing it on top of Windows.
I'm going to start selling (Score:5, Funny)
Re:I'm going to start selling (Score:1)
Re:WHOA (Score:3, Funny)
Re:No, but... (Score:1)
mirror (Score:3, Informative)
Re:mirror (Score:1)
Someone has an ego problem... (Score:5, Funny)
I know some insitutions rank their faculty based on the number of times their papers are cited, but they usually exclude self-citations in those counts.
Re:Someone has an ego problem... (Score:5, Informative)
When you're ahead of the crowd, you don't have many peers to cite.
Oh, and his reputation is actually pretty sound, he doesn't really need to rely on inflated citation counts, he has plenty of research dollars coming in -- that should keep his institutions (Argonne and U of Chicago) happy.
Re:Someone has an ego problem... (Score:1, Interesting)
Also as an FYI, these are bad papers to have on your C.V. unless you are the big person in the field. They amount to diddly squat unless you are the primary author or a bigwig.
Then again, I just might be jealous of the copious amount of citations but hey, whatever floats your boat I guess. Amen to being an anonymous coward....
Already happening (Score:5, Informative)
Re:Already happening (Score:5, Informative)
giFT/OpenFT tends to cater to the geek crowd right now since they don't release binaries, you have to understand CVS and keep up to date, and the best GUI is curses-based. That means you'll find more anime than Brittany Speers, but it's quick, reliable, and works great on Linux. (not affiliated, just a happy user
Re:Already happening (Score:2)
Re:Already happening (Score:1)
Re:Already happening (Score:1)
Re:Already happening (Score:2)
Re:Already happening (Score:1)
an even more daring prediction (Score:5, Insightful)
Here, I'll go out on a limb and make an even more daring prediction: grid computing will use Rendezvous-like services. Some of the machines may do that at boot time, to load customized and specialized machine configuration (you know, like BOOTP/DHCP followed by NFS), and others will use it at the application level to discover potential clients and servers.
All this stuff was designed into the Internet in decades ago. People are just giving fancy names to very traditional usages of sockets, servers, and broadcast packets. "Grid computing", too, is pretty much what people have been doing on networks of workstations for years: sometimes you push the jobs, sometimes available machines pull the jobs, sometimes you have a workflow manager, sometimes it's done through NFS, etc.
All of this reminds me of some teenager thinking that they are the first person on the planet to have discovered "sex".
Re:an even more daring prediction (Score:2)
when people started needing chat and file exchange servers because they weren't always on
On always-on workstations, this was traditionally handled via things like "talk" and "ftp".
I got rid of my old GRID computer years ago... (Score:5, Funny)
I predict. (Score:5, Funny)
Not long after that, the "virtual search warrant" and mandated backdoors in commercial grid products will appear.
Those fucks (**AA) would try to regulate telepathy if it existed.
-dameron
Re:I predict. (Score:1)
considering we pay for bottled water . . .
also the old MS joke that congressman made about charging for the alphabet and then charging an extra $40 (or whatever) if i decide to add letters such as N, or T
sorry, not that relevant, or funny
Re:I predict. (Score:3, Funny)
I realize the parent was using the asterisk as a wildcard, but as I was scrolled through the comments, it first appeared that he was willing to say 'fucks' but censored himself when referencing the RIAA, and the MPAA.
Good ol' unintentional comedy.
DC++ (Score:4, Interesting)
It's open source, and all, but there isn't a Linux client [sourceforge.net]. Any l33t coders out there that are bored should look at bringing this to the land of Linux.
And yes, I understand the irony of calling them OS-disabled, and in the same breath complaining that my OS of choice doesn't have the same facility.
Re:DC++ (Score:2, Interesting)
Don't know if this one is DC++ or just plain DC (what's the difference between them anyway?), but it has nice features like multi-hub search, and I've used it successfully.
I just feel the problem with DC is that there doesn't seem to be any way to resume broken downloads from other sources automatically (or maybe it's just this client).
Re:DC++ (Score:2)
Re:DC++ (Score:1)
dcgui offers the same functionality as DC++, with a few extras like download from multiple sources and upload speed throttling.
Since a large percentage of hub ops seems to be ignorant bastards who don't know how to read a dc client signature, these features have sadly led to dcgui being banned on many hubs. Of course, if the hub doesn't read the client signature at all, one can always turn off dcgui's signature and connect anyway, and hope that the ops don't stick their noses in your client.
Re:DC++ (Score:1, Offtopic)
Harumph (Score:5, Interesting)
In the future:
Computers will control our houses and your high definition television will be your main terminal
Someone will make a mobile phone that doesn't suck as a PDA (or a PDA that doesn't suck as mobile phone)
We'll all evolve more agile thumbs from "texting"
There will be One True programming language (not a troll)
Everyone will type on a Dvorak keyboard when not using a flawless voice interface that does what you mean and not what you say
We will bathe everyone in the electromagnetic glory of Wireless
As computers get faster and faster, and software gets more and more efficient, every user interaction will receive nearly instantaneous responses
VR is the next big thing
Build from a solid foundation and some of things will happen. Build from fragile abstractions and a sneeze will knock out the grid. The promise of technology is not the promise of earnings or market creation. How well does it help us live our lives.
Flush toilet, books == good
Pager, way-too-fast-food == bad
...Crawls back in cave...
If they do converge... (Score:3, Funny)
Either illegal or pointless (Score:3, Insightful)
As for "grid computing", if there was a real need for it, people who needed it would be buying up off-peak time on server farms. That's not happening.
Both ideas are promoted by people desperately seeking a revenue stream, rather than trying to provide a new capability. Unless they figure out some way to put a boot on the consumer's throat and make him pay, it's not going to happen.
Re:*Legal* trading evading extra-legal restriction (Score:3, Informative)
Re:*Legal* trading evading extra-legal restriction (Score:1)
ok back on topic, i think the whole media sharing thing could be less complex if the goverment would give the artist two options, one: this media that you have produced (name of media here) is free to reproduce and distribute aslong as no profits are made (except by the artist) for the material. two: this material is the property of the artist and cannot be distributed unless the artist receves compensation for every copy distributed... take out the middle man, the RIAA and the Music Lables... there old and pointless. well that is about all i can think of for now...
P2P Name Game (Score:2)
Even the definition of grid computing in this paper, is the same as distributed computing using a hybrid centralized+decentralized P2P topology.
"Invitation Only" (Score:3, Interesting)
duh (Score:3, Insightful)
I hope to see some of plan 9 [bell-labs.com] in "the grid". Need another CPU? Mount it into the filesystem...
Predicted?!?! (Score:3, Funny)
Re:Predicted?!?! (Score:1)
Re:This is why our government is going bankrupt. (Score:1)
ok, im gonna go join Jar-Jar Binks in pink spandex on his unicycle.
WHAT A COINCIDENCE!!! (Score:2)
Massively Parallel Games (Score:1)
Sony and IBM are throwing a lot of attention towards butterfly.net [butterfly.net].
It's a small company in WV, but with some interesting applications for grid computing with massively parallel games.
massively-multiplayer online games (MMOGs) (Score:1)
"Scooped, Again" says it better (Score:4, Interesting)
The paper's title refers to the Web having been implemented by those outside the systems research community, who had elegant solutions to interesting problems but didn't pay enough attention to needs of users. The authors are afraid this might happen again if P2P researchers ignore the needs of Grid users. The third generation P2P infrastructures represented by systems such as Tapestry, Pastry, and Chord are amazing. For example, with one of these, you could implement a truly distributed DNS system that doesn't use hierarchy or centralization, and thus would be much more immune to DoS attacks than the current system. P2P researchers should heed the Call to Action at the end of this paper.
Maybe not (Score:2)
Not merging, just sharing ideas (Score:1)
P2P Melded with NAT,litigation proof? (Score:1)
P2P Melded with NAT, is this a litigation proof model for P2P?
p r.php [eff.org]
. I propose a new P2P technology that mixes P2P and NAT technologies. Each client spends half of it's bandwidth on sharing it's own files, the other half on NATing other people's files. Direct connections to download a file are never permitted, all downloads must bounce through at least one NAT hop, and the clients never respond back to a search request with the location of the file, only that they know where to get the file from and the bandwidth that they can deliver the file at.
I having been thinking about the RIAA and MPAA hiring organizations to monitoring people's file sharing activities and then using that information to prosecute individuals. The recent Verizon case comes to mind http://www.eff.org/Cases/RIAA_v_Verizon/20030121_
In this type of system, which I'm sure can easily be designed and implemented, are the relaying PC's liable for the files they transmit, or do they fall under a "safe harbor" ISP type provision? Or is all this pointless and we should just push Congress for the type of "media tax" that is now placed on blank CD-Rs, tapes, etc to be applied to consumer bandwidth in exchange for a truce with the RIAA, MPAA, etc?