Sep 13 12:32:32 ok, Norm; I see TAG_Weekly()12:30PM scheduled to start 2 minutes ago Sep 13 12:32:52 --- Norm has changed the topic to: http://www.w3.org/2001/tag/2005/09/13-agenda.html Sep 13 12:32:58 --- Norm has changed the topic to: TAG: http://www.w3.org/2001/tag/2005/09/13-agenda.html Sep 13 12:34:22 Meeting: W3C TAG telcon Sep 13 12:34:27 Chair: Vincent Sep 13 12:34:32 Scribe: Norm Sep 13 12:34:37 ScribeNick: Norm Sep 13 12:35:49 Regrets: TimBL, HT, NM, Roy Sep 13 12:35:50 IRC: http://www.w3.org/2005/09/13-tagmem-irc Sep 13 12:57:27 --> Vincent (quint@194.199.20.105) has joined #tagmem Sep 13 12:58:33 * Norm waves to Vincent Sep 13 13:00:48 --> Ed (chatzilla@15.244.168.93) has joined #tagmem Sep 13 13:01:17 TAG_Weekly()12:30PM has now started Sep 13 13:02:04 +Vincent Sep 13 13:02:44 +Norm Sep 13 13:03:35 zakim, [INRIA is Vincent Sep 13 13:03:35 +Vincent; got it Sep 13 13:03:39 zakim, who's on the phone? Sep 13 13:04:14 * DanC Zakim, call DanC-BOS Sep 13 13:04:15 * Zakim ok, DanC; the call is being made Sep 13 13:04:16 +DanC Sep 13 13:04:38 zakim, who's here? Sep 13 13:04:38 On the phone I see Ed, PaulStrong, Vincent, Norm, DanC Sep 13 13:04:39 On IRC I see Ed, Vincent, RRSAgent, Zakim, Norm, DanC Sep 13 13:05:01 Agenda: http://www.w3.org/2001/tag/2005/09/13-agenda.html Sep 13 13:05:11 Topic: Administrivia Sep 13 13:05:45 Most of today is for GRID discussions Sep 13 13:05:59 Accept minutes of last telcon: http://www.w3.org/2001/tag/2005/09/06-minutes.html Sep 13 13:06:18 Accepted (Vincent will remove "DRAFT"). Sep 13 13:06:31 Topic: Discussion of GRID Sep 13 13:06:38 Thanks to Paul Strong for joining us Sep 13 13:06:52 This is an informal discussion of GRID and it's connection to the Web Sep 13 13:08:50 Paul: Paul Strong is a Systems Architect at Sun. Works in the N1 product group. N1 is a suite of products that leverage the GRID Sep 13 13:09:08 ...Grid is a somewhat ambiguous term being widely used by vendors Sep 13 13:09:31 ...Within N1, I've been working on products for about five years. Mostly working on data center and enterprise applications Sep 13 13:10:16 ...Recommends July issue of ACM Queue Sep 13 13:10:17 * DanC hopes for an answer to "what is the grid?" by value, not just by reference Sep 13 13:10:57 ...GRID is a view of the networking infrastructure Sep 13 13:11:15 ...It's a view of computing resources that are pervasive. It's more about the platform than the end-user applications Sep 13 13:11:28 (hm... http://www.sun.com/software/gridware/index.xml Sun N1 Grid Engine 6 ... seems to be a hunk of hardware. I thought maybe N1 was a service.) Sep 13 13:12:07 ...GRID really is about recognizing two trends: growth in network bandwidth, and network distributed services Sep 13 13:12:36 Paul: GRID platform offers scalability, redundancy, ... Sep 13 13:13:35 Paul: Needs services for distributing and managing work loads Sep 13 13:13:54 ...Analogous to an electrical grid, in the sense that it's pervasive and more-or-less uniform Sep 13 13:14:01 (hmm... http://en.wikipedia.org/wiki/Grid_computing "The SETI@home project, launched in 1999, is a widely-known example of a simple grid computing project." ) Sep 13 13:14:11 +DOrchard Sep 13 13:14:34 DanC: Sun N1 Grid seems to be a hunk of hardware... Sep 13 13:14:41 --> dorchard (3f60a3ac@128.30.52.23) has joined #tagmem Sep 13 13:14:46 Paul: The N1 products are a mixture of both hardware and services Sep 13 13:15:09 Paul: Software is a meta-operating environment. Those products are called N1 Sep 13 13:15:42 ...They're closely tied to a set of hardware to run them on at Sun. The result is an integrated set of components. You no longer care about individual servers or OS instances. Sep 13 13:15:54 DanC: So if I buy a chunk of N1, do I get CPU hours or a box? Sep 13 13:16:23 Paul: It depends what you want, you can buy time on our GRID, or buy hardware and setup your own Sep 13 13:16:56 Paul: An example of a GRID application is SETI@Home Sep 13 13:17:17 ...The use of the term GRID was prevalent initially in scientific and academic community. Sep 13 13:17:28 ...In the commercial space, rendering and simulation applications Sep 13 13:17:49 ...The software that allows that workload to be distributed/managed/aggregated is the middleware, integration layer that is the meta-operating environment Sep 13 13:18:26 DanC: Is it a style of computing, or is it technical standards that you could interoperate with? Sep 13 13:18:29 Paul: It's some of both Sep 13 13:18:49 DanC: Does SETI@Home conform? Sep 13 13:18:56 Paul: No, it predates them. The context is still being refined. Sep 13 13:19:11 Paul: There are a couple of consortia working on this: The Global Grid Forum Sep 13 13:19:28 ...There's The Enterprise GRID Alliance, focused on driving GRID adoption within enterprises Sep 13 13:19:45 (http://en.wikipedia.org/wiki/Grid_computing doesn't seem to mention The Enterprise GRID Alliance ) Sep 13 13:19:52 Paul: To get the GRID used in less compute-intensive environments Sep 13 13:20:09 -> http://www.gridalliance.org/ Enterprise GRID Alliance Sep 13 13:20:49 Paul: discusses benefits of GRID: ability to manage pools of resources; a mutable, dynamic space Sep 13 13:21:42 Paul: reiterates the goal of treating these things holisticly... Sep 13 13:22:18 (EJB and J2... missed. hmm... I was starting to understand...) Sep 13 13:22:20 Paul: workload management, mechanisms for monitoring, managing, controlling processes Sep 13 13:23:02 Paul: Users need to be able to combine a heterogeneous set of products and services together Sep 13 13:23:17 Paul: Standards are needed to allow each of these components to be managed. Sep 13 13:24:18 Paul: The term GRID has become very loaded. Sep 13 13:25:23 [scribe lost thread] Sep 13 13:25:55 There's lots of marketing in this space: managing complexity, providing agility, etc. Sep 13 13:26:21 Paul: They're very similar, but they aren't identical. The GRID space is very confusing for many of the end-users and consumers. Sep 13 13:26:45 Ed: GRID is a very broad term. Everything from SETI@Home to shared system resource pools that's more of a realtime virtual machine type of thing Sep 13 13:26:48 Paul: Yes, absolutely. Sep 13 13:27:01 Paul: One of the difficulties we have as an industry is articulating this Sep 13 13:27:43 ...It's going to take a long time to get to the end. Sep 13 13:28:35 Paul: A lot of the technologies we think about today in the GRID space that do the mapping of workload onto resources Sep 13 13:29:08 Paul: There are also provisioning services Sep 13 13:29:43 Paul: What we're automating today is the provisioning processes, but that's just the beginning. Sep 13 13:29:48 DanC: How is provisioning expensive? Sep 13 13:30:10 Paul: Consider an electronic book store that has a web tier, a web service tier(?), and a database server tier Sep 13 13:30:33 Paul: There's a set of database servers running on particular Sun hardware with a particular OS Sep 13 13:30:47 Paul: The services layer might be BEA running on some particular Dell hardware Sep 13 13:30:59 ...Right now there isn't a standardized way to describe all these components Sep 13 13:31:18 ...Not only are the components complex, but there's a relationship with every other component already in the data center Sep 13 13:31:35 ...Today, people manage individual resources Sep 13 13:31:41 ...But those are increasing exponentially Sep 13 13:31:56 ...Because they don't trust management tools, each server is typically dedicated to a single function Sep 13 13:32:11 ...This leads to silos of services that perform single tasks Sep 13 13:32:19 ...This leads to waste and lack of agility Sep 13 13:32:29 ...It's very hard to track relationships between all the components Sep 13 13:32:48 DanC: Are there any GRID computing saves the day stories? Sep 13 13:32:57 Paul: There are stories that it's leading that way Sep 13 13:33:33 ...A lot of stuff is relatively static today. We have a tool that allows you to provision complete projects, like the bookstore Sep 13 13:33:42 ...It does all the work Sep 13 13:34:06 ...It typically pays for itself in six to twelve months. There are fewer unplanned outages because planned downtime is all automated Sep 13 13:34:17 ...It's more deterministic in production and is more reliable. Sep 13 13:34:39 Paul: The developers can create the model when they create the application. For provisioning the test and QA engineers can test with a single button. Sep 13 13:35:00 DanC: It has a little blinking light that says "you need a new database server" Sep 13 13:35:08 Paul: Yep. Sep 13 13:35:30 [Scribe hears something about ad hoc construction that seems at odds with the previous story..] Sep 13 13:35:52 Paul: When load gets high, the provisioning application will attempt to reconfigure (scribe ?) Sep 13 13:36:31 Paul: Getting to the point where it all "just works" is going to take a long time. It's very easy to solve problems with regards to concrete things, but it's far more complicated when you're trying to model more abstract components (a server vs. a tier of servers) Sep 13 13:37:24 DanC: It's all proprietary things cobbled together, but Sun does have products in this space? Sep 13 13:37:38 Paul: Yes. It's mapping workload onto resources with respect to policy. Sep 13 13:38:17 HP and IBM do as well. Unfortunately, they don't work together to create one grid, each has its own grid. Sep 13 13:38:18 Paul: In the GRID world, we're talking about mapping services (a bookstore, SETI@home, etc.) onto a network of resources (servers, firewalls, etc.) with respect to policies Sep 13 13:38:38 Paul: The first things that get automated are the simple mechanisms. Sep 13 13:39:10 Paul: There will eventually be a move towards automating higher order problems, such as managing performance and availability. Sep 13 13:39:35 Paul: Today there are no single products that let you do all of those things Sep 13 13:39:56 ...Instead you get different products to manage different aspects of that. You get something that is more automated, but still has lots of human interaction Sep 13 13:40:19 Paul: Sun has products that fit into a number of those spaces, but none are integrated together as a whole meta-operating system. No one's products are. Sep 13 13:41:26 Vincent: What are the consortia doing today, what are the main standards under development? Sep 13 13:41:49 Paul: Several things are needed Sep 13 13:42:09 ...A way of describing the requirements of the system Sep 13 13:42:38 The Enterprise Grid Alliance is working on this sort of thing Sep 13 13:42:43 q+ to ask if these enterprise grids have peers grids Sep 13 13:42:43 * Zakim sees DanC on the speaker queue Sep 13 13:42:46 Paul: And use cases based on that description Sep 13 13:43:52 Paul: We're working on a standard set of requirements that we can give to other standards organizations Sep 13 13:44:04 Paul: The Global Grid Forum is working on standards farther downstream Sep 13 13:44:20 ...A service-centric architectural view; the OGSA (Open Grid Services Architecture) Sep 13 13:45:03 Paul: Because GRID was originally driven by compute-intensive applications, they have a lot of those, but they're working on getting more broad Sep 13 13:45:27 Paul: A job control language is one example. How do I describe a work load, schedule it, monitor it, etc. Sep 13 13:45:58 Paul: As you approach the more concrete things, you want to standardize them too. That's where interaction with DNTF occurs. Sep 13 13:46:11 DMTF = Distributed Management Task Force (www.dmtf.org) Sep 13 13:46:14 s/DNTF/DMTF/ Sep 13 13:46:22 They own the SIM standard (Standard Information Model) Sep 13 13:46:50 * DanC was going to say all the task forces he knows are parts of other things, but realizes the IETF is counter-example #0 Sep 13 13:47:16 There's work to make some of these things more abstract as well (pools of servers instead of single servers) Sep 13 13:47:29 Paul: There are OASIS GRID/WS standards under development as well Sep 13 13:47:59 Paul: You can look at GRID as the platform that is the network that is the web Sep 13 13:48:17 Paul: There are other standards in this space too (for storage, for example) Sep 13 13:48:42 ack danc Sep 13 13:48:42 DanC, you wanted to ask if these enterprise grids have peers grids Sep 13 13:48:44 * Zakim sees no one on the speaker queue Sep 13 13:48:54 DanC: Are enterprise grids mostly their own world, or do they have peers? Sep 13 13:49:00 DanC: Does my grid talk to other grids? Sep 13 13:49:34 Paul: We define an enterprise grid as the set of components (from disks to CRM applications) managed by a single enterprise Sep 13 13:49:47 Paul: But each may have several data centers Sep 13 13:50:17 Paul: In some sense, they're isolated in terms of management, but they do interact with the Web. Sep 13 13:50:40 Paul: And one enterprise grid could interact with another (the bookstore grid interacting with the credit card company grid) Sep 13 13:50:58 DanC: How will these two talk to each other? Sep 13 13:51:19 Paul: The expectation is that we'd be using standard mechanisms for interaction Sep 13 13:51:40 Paul: But I as the bookstore owner may have expectations about the speed of service from the credit card company Sep 13 13:51:52 ...I may want to negotiate that quality of service. Sep 13 13:51:58 ...Possibly on a per-transaction basis. Sep 13 13:52:27 If my customer is a real brick-and-mortar store ordering thousands of books, I may want a faster answer than for Joe Individual User. Sep 13 13:53:17 Paul: We chose to bound the problem at a single enterprise because it makes authority and control simpler Sep 13 13:53:37 Paul: When you're working across enterprises, then you have federation rather than hierarchy Sep 13 13:54:25 Paul: GGF views its charter as everything grid, they see what EGA does as (an important) subset Sep 13 13:54:54 Paul: They care about viewing the internet as a set of computers controlled by different organizations but on which I could impose a virtual organization Sep 13 13:55:09 q+ to ask about job migration between, say, sun's and IBM's grid services Sep 13 13:55:09 * Zakim sees DanC on the speaker queue Sep 13 13:55:28 Paul: For example, automobile design is sometimes shared across companies because it's so expensive Sep 13 13:55:52 Paul: From the GGF perspective, a virtual GRID could be constructed between these companies Sep 13 13:56:14 Paul: Typically, the shared resources are segregated from the companies own resources Sep 13 13:56:17 q? Sep 13 13:56:17 * Zakim sees DanC on the speaker queue Sep 13 13:57:42 Ed: It seems like because the GRID is undefined, a lot of work is hindered. If it's more along the lines of a distributed computing environment, then I can see where that comes into play. Is there progress on defining either striations or a clear definition of what GRID is? Sep 13 13:57:50 Paul: In terms of the word GRID, no Sep 13 13:58:44 Paul: We're working on this to some sense in EGA by working on requirements. By being able to clearly enumerate and describe problems, we can guide GGF to work on a particular area. Sep 13 13:59:29 Paul: A big challenge is identifying the set of problems that people care about most and the boundary between the components we care about. Sep 13 14:00:05 Paul describes a number of things that can be virtualized Sep 13 14:00:26 Paul: Having a model for these components and the life cycle of those components is critical for the standards bodies to be able to do stuff that isn't unintentionally competitive Sep 13 14:01:04 Ed: Right, and I guess that's why I think breaking the big problem down into smaller problems seems like something you'd want to do Sep 13 14:01:38 Paul: GGF is more of a boil the ocean perspective, EGA is about boiling enough water to make a cup of tea Sep 13 14:02:05 Paul: There is a working group called the SCRUM (scribe wonders about spelling) in GGF that's trying to look at these issues Sep 13 14:02:56 ack danc Sep 13 14:02:56 DanC, you wanted to ask about job migration between, say, sun's and IBM's grid services Sep 13 14:02:58 * Zakim sees no one on the speaker queue Sep 13 14:03:21 DanC: If Amazon rented time on the Sun N1 thingy and some IBM On Demand computing, is it feasible to migrate jobs across those? Sep 13 14:03:29 Paul: It totally depends. Sep 13 14:03:47 Paul: There are certain classes of workflow where you can migrate the work today. In a batchable system, you could move them around in stages. Sep 13 14:04:21 Paul: Rendering would be a good example. I've got 20,000 jobs, I can send 10,000 to each. 3,000 fail on one system so I can migrate them to the other. Sep 13 14:05:08 Paul: If you have shared infrastructure, you can migrate between transactions Sep 13 14:05:24 DanC: Across the Sun/IBM boundary? Sep 13 14:05:39 Paul: Technically, yes. Sep 13 14:06:07 Paul: Right now a lot of this is really proprietary. It'll become easier after the standards are written. Sep 13 14:07:28 Paul: People are mainly looking at whole data centers or whole enterprises at the moment. Sep 13 14:08:18 Vincent: Is there anything important that you feel wasn't addressed? Sep 13 14:08:25 Paul: I'm not really sure. Sep 13 14:08:51 Paul recommends ACM Queue Magazine again Sep 13 14:09:38 Most of the articles will be online soon. Sep 13 14:09:46 http://www.acmqueue.org/ Sep 13 14:10:13 TAG thanks Paul for a great overview. Sep 13 14:10:31 * DanC made it as far as http://www.acmqueue.org/ , but doesn't know what link to follow next Sep 13 14:11:03 Vincent: Thanks also to Norm for organizing Sun's participation Sep 13 14:11:09 -PaulStrong Sep 13 14:11:13 Norm: Thanks again, Paul Sep 13 14:11:20 Topic: Edinburgh Face-to-Face Sep 13 14:11:47 Draft agenda: http://www.w3.org/2001/tag/2005/09/20-agenda.html Sep 13 14:12:07 q+ to ask for abstractComponentRefs-37 on the ftf agenda, maybe Sep 13 14:12:07 * Zakim sees DanC on the speaker queue Sep 13 14:12:12 Vincent: Some time for issue status, then time for four or five issues to discuss. Sep 13 14:12:21 Vincent: Return to the discussion of new directions. Sep 13 14:13:17 ack danc Sep 13 14:13:18 DanC, you wanted to ask for abstractComponentRefs-37 on the ftf agenda, maybe Sep 13 14:13:21 * Zakim sees no one on the speaker queue Sep 13 14:13:47 DanC feels more prepared to talk about abstractComponentRefs-37 Sep 13 14:14:29 Vincent: Try to review the draft agenda over the next day or so and send feedback so it can be updated before the f2f. Sep 13 14:14:36 Vincent: Any other business? Sep 13 14:15:23 * Norm notes that IRC will at least be available Sep 13 14:16:45 -DOrchard Sep 13 14:16:45 Next meeting is the f2f on 20 Sep in Edinburgh Sep 13 14:16:52 -DanC Sep 13 14:16:56 -Norm Sep 13 14:16:57 Adjourned Sep 13 14:16:58 -Ed Sep 13 14:16:58 -Vincent Sep 13 14:16:59 zakim, bye Sep 13 14:16:59 <-- Zakim (rrs-bridgg@128.30.52.30) has left #tagmem Sep 13 14:17:00 TAG_Weekly()12:30PM has ended Sep 13 14:17:02 Attendees were PaulStrong, Vincent, Norm, Ed, DanC, DOrchard Sep 13 14:17:05 <-- Ed has quit (Quit: Chatzilla 0.9.65 [Mozilla rv:1.7.5/20041107])