Convergence of P2P and Grid Predicted

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...

[ObviousGuy]

Sounds like the P2P folks are getting a little antsy looking for any evidence that P2P isn't just a really good way to encourage copyright infringement.

Grid computing can survive just as well without P2P. I'm not so sure that it's the same in reverse.

Grid Computing,

[Visaris]

As my professor described it, is a system similar to a power grid. You can plug in anywhere, and use the resources (Disk, Memory, CPU) of the grid for computation. Your resources would be added to the "grid collective" as well.

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.

an even more daring prediction

[g4dget]

No matter what happens, these people can claim victory. I mean, P2P basically could be anything where one machine sometimes acts as a client and sometimes as a server. You know, like workstations used to before the advent of dial-up Internet access, when people started needing servers because they weren't always on or their IP addresses kept changing. And for grid computing, well, machines sometimes need to act as clients and sometimes as servers.

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".

Seti@home

[bigsexyjoe]

What about the Seti P2P system. Who does that steal from? Well, what can we expect since you clearly commented before having time to read the story? Just so you know the story's focus was more techinical and it talk about what a P2P network is abstractly. It discussed what it could be when combined with Grid Computing.

It is a serious well supported argument. You are just shooting your mouth off with ad hominiem attacks which probably aren't valid.

Actually

[ObviousGuy]

If you redefine anything that you like into terms that you find useful, then you can make your argument look really good.

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:Grid Computing,

[smd4985]

Some current Gnutella clients could do grid computing pretty easily - I think the question is demand. LimeWire uses Java so technically one could create a 'JobInterface' that could be divided amongst peers for execution (definitely abstracting some issues). The big problem is that the common user doesn't have the need to write programs that need help from disparate peers. Not until there is a very high-level programming language that Joe User could make effective use of AND Joe User has a need for a lot of CPU cycles will Grid Computing features in P2P networks make sense.

Then again, perhaps it is a case of "If you build it, they will come."

Either illegal or pointless

[Animats]

The whole "peer to peer" network thing exists solely to evade copyright. If it wasn't for that, music would just go out on USENET, which is far more efficient. Each new item would traverse each link no more than once, and usually less. Less-used stuff would be downloaded from web servers. How many new mainstream songs come out every day, after all?

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.

duh

[lightray]

Convergence of P2P and Grid?? Well, DUH. They're basically two names for the same thing. Almost.

I hope to see some of plan 9 [bell-labs.com] in "the grid". Need another CPU? Mount it into the filesystem...

Re:Sounds like...

[JMZorko]

I've been involved in a lot of P2P work, and it is definitely a wonderful technology, and it also solves a big problem -- it is an enabler for reaching ever-larger numbers of people / clients. P2P systems can be implemented such that the more clients participate, the more room there is for clients to participate. Like the article says, P2P solves the "my parent node failed" problem pretty well (by dynamically reconnecting to another parent node).

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:TV and Computers will converge

[Planesdragon]

Er, they have.

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.