Access Grid Meeting 10/02/02 Attendees: Manchester - Mike Daw Boston U - Jennifer NewMIC - Brian, Hanyi Vrije U - Luke U Buffalo - U Montana - Don Morton Dartmouth - SDSC - mgates Sydney - North Dakota State U - CHPC - UTAH Network Services Mike Daw: Why the differentiation between service types? Ivan: Conceptually and architecturally, it is helpful. We have discussed Brian/NewMIC: AG1 was a discrete set of known things like audio, video, etc., AG2 will be a flexible extensible system. Do you have a plan for managing this? Ivan: ... NCSA: What if I walk in to a meeting with my laptop and a gizmo, and want to make my gizmo available in the venue? Do I have to write code, or can I just fill out a form on a web page, or? Ivan: If your gizmo is networked, you may have to write the code for a service to make the gizmo available. We will optimize this over time, to minimize the effort required to connect gizmos. Voyager Jennifer: People are always wanting to explore voyager recordings into mpeg, etc. Any thoughts on that? Ivan: What we're promising is to be able to export the data in a well documented form, to enable third parties to develop converters. We want to allow people to take Voyager recordings and do whatever they want with it. That is a promised deliverable. We'll give examples on how to read and write the data structures Jennifer: With recordings becoming more reliable, is there a legal issue? Is anyone talking to an attorney? Ivan: We're avoiding talking to lawyers on this subject. The law is strange in dealing with patents. We see a need to be able to ask people for authorization to record, where if they say no, voyager would drop their stream. If people are properly identified, we will be able to reocrd the authorization given by a person for later proof. Jennifer: For now, we are using the clinet that you let us vnc to. What is the plan for a clinet? Ivan: Could be Java, Quicktime, other. We will probably not ever have a Windows media framework because they use their own protocol instead of RTSP. When we get to non-streaming media, the other clients will fall over. So we'll have to write a new client. There may be two clients, which may or may not be glued together. Jennifer: Real player can't be used for browsing voyager archives Ivan: Right. We will develop a browsing client, which should be integrated with playback tools to launch them with the selected session. Another possibility is a web page that describes sessions and how you can play them back in your particular software. We will try to avoid having the clinet vnc to the server. ag-tech2-evl:Is there any easy way to grab the raw voyager recording data to play back in locally (doesn't have to be mpeg or other exported format)? Ivan: That is exporting. We don't have any mechanism for going around the interface and getting the raw data. That would leaad to chaos and mayhem for users and developers. So the only mechanism is through exporting. Jennifer: What are we doing in 2.0 to address issues with multicast and networking? (this particular question and answer were horribly plagued by network failures) Ivan: Services will monitor network activity and try to automatically heal through bridging. We want to get the tools in place and then do algorithmic research like how to automate bridging. For example, take advertising of a bridge: Sure, someone may be offering bridging, but you know you don't want to bridge with them for some reason. Or, imagine a geographically defined bridging tree, which would specify useful bridges between sites. Jennifer: We were chatting about getting requirements from people who plan events of various kinds. I think those people would not be interested in understanding the technical underpinnings of the ag, but maybe you could do a talk for people planning events to describe what they can expect in ag2. Ivan: That's a good idea. We could solicit things from ag-tech, and discuss these issues with people to get feedback. The answer to your question is yes, we could do a boiled down version of these talks to node managers.