CURRENT_MEETING_REPORT_ Reported by John Johnston/National Semiconductor Minutes of the IP Routing for Wireless/Mobile Hosts Working Group (MOBILEIP) Agenda o Review of mobility model o Liaison reports o Document status o Subcommittee reports o Short presentations o Interim Meeting Mobility Model Greg Minshall reviewed the mobility model for the first time attendees in the session. Basically, the problem was stated as finding a methodology (architecture, protocol, etc.) to support the routing of mobile hosts (MH). In most of the models presented to date, each MH has both a home address and a forwarding or care of address. Packets are sent to a MH via its home address. These packets are directed via normal routing to a base station which serves as the home base of the MH. This base station must know where the MH is at all times by maintaining the IP address of the base station currently serving the MH. Assuming the MH is ``not at home'' the base station then forwards the packet to its peer currently serving the MH, who in turn delivers the packet directly to the MH. The base station to base station delivery mechanism is called tunneling which can result in inefficient (dogleg) routing. An extreme example is a US-based Amsterdam IETF attendee trying to connect to a local Amsterdam host. Packets would be first routed over the Atlantic to the home base station in the US, then routed back over the Atlantic to the MH's current base station in Amsterdam, and finally delivered to the local Amsterdam host. This problem will be handled through the use of address caching. Finally, Greg clarified the scope of the working group as supporting media independent mobility. One solution must handle wireless IR, wireless RF, ethernet, etc. Liaison Reports Charlie Perkins reported that 802.11 is standardizing IEEE MAC protocols for wireless media. This body is meeting during the same week as the IETF. At their last session, Charlie proposed that the IETF working group inform 802.11 about all of the network layer events and indications that will be necessary to support mobile IP. Charlie indicated that 802.11 still has many open issues including MAC address selection (48-bit?). Steve Alexander did not report on the Dynamic Host Configuration Working Group (DHC) because the group has not met since the last IETF. Scott Kaplan, the liaison for the Domain Name System Working Group (DNS), did not report, because he was unable to attend the last DNS session. There was no report given for the Internet Protocol Security Protocol Working Group (IPSEC) because John Ioannidis was unable to attend this MOBILEIP session. Document Status o Fumio Teraoka from Sony will distribute a new version of his document when he returns from the IETF. o Dave Johnson has an updated version of his document available. o The working group has not heard from Columbia University. o Charlie Perkins will complete a new version of his document in several months. o Matsushita's draft work is continuing. A version will be released in several weeks. Subcommittee Reports o Terminology (Perkins) Charlie conveyed the confusion over inconsistent terminology by offering the following list: Host Mobile Host Correspondent Host - used to describe a destination host which may or may not be mobile Home Address Agent/Home Redirector (Location Server) Location Directory Base Station (Cell Manager) Visited Redirector/Remote Redirector Visitor's Redirector/Vistor's Redirector Cell Cluster Consistent terminology has not yet been selected. Consensus was that this must happen as soon as possible, but it was recognized that we must agree on the functionality of the things we are naming before we can agree on their names. Acronyms should be avoided at least until consistent terminology is adopted. o Beaconing (Perkins) Charlie described beaconing as the method by which a Mobile Host advertises itself to a base station. Some link layers have beaconing built-in and will not require layer 3 support. Mechanisms must be identified to support advertisement requests (solicitation), acceptance, and responses. Broadcast and multicast mechanisms are both under consideration. Open question - Can we use existing router discovery methods? o User Requirements (Rehkter) Yakov presented an overview of his latest document which will be available in several weeks. This document addresses the Internet model, mobile computing models, mobility dynamics, overhead considerations, scalability, performance transparency, manageability, security, and a comprehensive solution. o Database Service Interface (Rehkter) - no report o Tunneling/Redirecting (Dave Johnson) Tunneling is initiated by a Location Cache or a Location Server (Dave's terms). Dave is no longer supporting loose source routing because it is slow, buggy, and politically incorrect. Instead, the approach is to use encapsulation as described by Dave's latest document. This approach supports all existing IP options, cache coherency through the dynamic maintenance of an address forwarding list in each tunneled packet, loop discovery through the same address list, and base station state recovery. Cache coherency is also supported through a new mobile ICMP redirect message. Open questions are: 1. Does this break trace route? 2. Does this break MTU discovery? o Security (Ioannidis) - John Ioannidis was unable to attend o MIB - not discussed Short Presentations o Dave's Latest Document (Dave Johnson) Dave presented the essence of his latest proposal when he discussed tunneling/redirecting. It was decided to table further discussions until the WG had a chance to read the new version. o Review of Existing Mobile Services and Techniques (John Penners) John listed existing technologies and services. He briefly described RAM Mobile Data, ARDIS, PRMA (Packet Reservation Multiple Access), CDPD (Cellular Digital Packet Data), AMPS (Advanced Mobile Phone System), GSM (Groupe Special Mobile), and DECT (Digital European Cordless Telecom). See his document for for more details. o Experience with CU Mobile Implementation (Jim Binkley) Jim described problems encountered while porting the CU code. The 20 MHs were DOS based, while the 7 MSRs were ported to UNIX and supported mobility on each port. His results are based on 1 month of testing. - DOS MH implementation difficulties outline the need for and MH architecture ASAP. - CU requirements = ARP, route, and raw sockets. Needed help from FTP software. - A multi-home bug was discovered. Because forwarding address != IP source address, MSRs got into trouble looking at addressing in previous layers. - intra segment routing needs to be addressed (ARP) - problems with traceroute, IP options, and 8k fragmentation when using IPIP encapsulation - implementation bug in MSR led to infinite ack loops - Installation of JI's embedded subnet demonstrated importance that home subnet not be in too many (> 4) partitions. - JI's embedded subnet causes MSRs to burn packets with proxy ARP turned on. - directed broadcast caused broadcast storms - routers don't realize its a broadcast, so it is forwarded. - when Jim "unplugged it and watch"ed, PC based NFS crashed and BSD NFS/UDP exhibited slow start o Beaconing Procedure (Charlie Perkins) Charlie outlined a beaconing procedure: 1. MH solicits to either a known MAC (unicast) or an unknown MAC (broadcast/multicast) 2. IAP (Internet Access Point) advertises (unicast or broadcast/multicast) 3. MH sends acceptance (unicast, with encrypted value?) 4. IAP sends ack or reject 5. MH updates old IAP 6. MH updates home redirector and proposed a beacon packet format: type, code, checksum IAP address timestamp/serial # beacon interval/lifetime random value to by encrypted MAC address type of authentication (including none) The status of Charlie's work: - intend to try icmp "host moved" function in IAPs - intend to release code fragments - will switch from simple beacon - need to select encapsulation/get protocol #/modify IPIP - intend to support public domain Mach implementation - desires routing to "shared media" - port to OS/2 Interim Meeting There was consensus to meet somewhere on the east coast sometime before the November IETF. Possible dates and locations will be discussed via the mailing list. Further Discussion Yakov Rehkter's subcommittee report on user requirements created a discussion regarding CDPD. This technology was described by Mark Knopper: o Consortium of 9 large US and 1 Canadian voice carriers. o Data services over cellular infrastructure. o Mobile End System makes itself known to a Mobile Intermediary System. o Packets routed first to Intermediary System which forwards them to End System. o Billing through X.400. Suzy Brown expressed the desire for the IETF to press ahead quickly to avoid the potential for deployment of technology-specific solutions that will not interoperate with the Internet. Other infrastructure-based solutions are being developed (Ram Mobile Data, Mobitex, GSM, etc.). Along the same lines, John Penners' review of mobile services spawned discussions centering on the relationship between mobile IP and the many specialized services and providers. Steve Deering presented a model that emphasized the logical separation from wireless service providers and the Internet. This led to several observations: o We should view the technologies as physical medium below IP (service provider is at lower level). o Successful mobile IP deployment could leverage incorporation into technology-specific switches. o A goal is to avoid ``doglegs.'' o Single hop at IP level (not single physical hop) in service provider. o Understanding providers' rules might help the dogleg problem. o Service providers need mobile IP to connect their clouds. Greg Minshall presented a similar model emphasizing how CDPD could create a coast-to-coast dogleg because it is not care of address-aware. This led to a discussion over whether it would be beneficial to take proactive measures to influence CDPD. During Charlie Perkins' presentation on beaconing procedures, Steve Deering emphasized the desirability that mobile hosts transmit new base station updates (as opposed to IAPs). Also, Steve stated that he would like to use multicast addressing over broadcast whenever possible (addressing must be consistent within a cell), and Greg indicated that we should request a ``well known'' address for this purpose. Open Issues o Can we use existing router discovery methods to support beaconing? o Does Dave Johnson's encapsulation technique break traceroute? o Does Dave Johnson's encapsulation technique break MTU discovery? Action Items o Steve Deering will contact Columbia University for an update on their work. o Determine if a working group should be formed within the IETF to deal with the issues of encapsulation. o Obtain a well known multicast address. o Obtain new ICMP number. Attendees Kannan Alagappan kannan@DSMAIL.ENET.DEC.COM Steve Alexander stevea@lachman.com Nick Alfano alfano@mpr.ca James Allard jallard@microsoft.com Bernt Allonen bal@tip.net Michael Anello mike@xlnt.com Anders Baardsgaad anders@cc.uit.no Cynthia Bagwell cbagwell@gateway.mitre.org Dennis Baker dbaker@wellfleet.com John Ballard jballard@microsoft.com Nutan Behki Nutan_Behki@qmail.newbridge.com Per Bilse bilse@ic.dk Jim Binkley jrb@ibeam.intel.com Carsten Bormann cabo@cs.tu-berlin.de Michael Brescia Ronald Broersma ron@nosc.mil Ramon Caceres ramon@mitl.research.panasonic.com Thomas Cordetti tomc@digibd.com Geert Jan de Groot geertj@ica.philips.nl Stephen Deering deering@parc.xerox.com Pierre Dupont dupont@mdd.comm.mot.com Kjeld Borch Egevang kbe@craycom.dk Julio Escobar jescobar@bbn.com Dan Frommer dan@jeremy.enet.dec.com Shoji Fukutomi fuku@furukawa.co.jp Robert Gilligan Bob.Gilligan@Eng.Sun.Com Jari Hamalainen jah@rctre.nokia.com Mark Handley mhandley@cs.ucl.ac.uk Gerd Holzhauer Holzhauer1@applelink.apple.com John Hopkins J_Hopkins@icrf.icnet.uk Steven Horowitz witz@chipcom.com Chris Howard chris_howard@inmarsat.org Geoff Huston g.huston@aarnet.edu.au David Johnson dbj@cs.cmu.edu John Johnston john@berlioz.nsc.com Philip Jones p.jones@jnt.ac.uk Marijke Kaat marijke@sara.nl Scott Kaplan scott@wco.ftp.com Ton Koelman koelman@stc.nato.int Mark Kosters markk@internic.net Mark Laubach laubach@hpl.hp.com Tony Li tli@cisco.com Susan Lin suelin@vnet.ibm.com Cynthia Martin martin@spica.disa.mil Donald Merritt don@arl.army.mil Paul Milazzo milazzo@bbn.com Greg Minshall minshall@wc.novell.com William Miskovetz misko@cisco.com Keith Mitchell keith@pipex.net Henri Moelard henri.moelard@utrecht.ncr.com Jun Murai jun@wide.ad.jp Ronny Nilsen Ronny.Nilsen@usit.uio.no Petri Ojala ojala@eunet.fi Zbigniew Opalka zopalka@agile.com John Penners jpenners@advtech.uswest.com Charles Perkins perk@watson.ibm.com Jim Rees Jim.Rees@umich.edu Yakov Rekhter yakov@watson.ibm.com Henry Sanders henrysa@microsoft.com Hal Sandick sandick@vnet.ibm.com William Simpson Bill.Simpson@um.cc.umich.edu Fumio Teraoka tera@csl.sony.co.jp Antoine Trannoy trannoy@crs4.it Thierry Turletti turletti@sophia.inria.fr Werner Vogels werner@inesc.pt Jost Weinmiller jost@prz.tu-berlin.d400.de Kirk Williams kirk@sbctri.sbc.com Rachel Willmer rachelw@spider.co.uk Sam Wilson sam.wilson@ed.ac.uk Wilfried Woeber Wilfried.Woeber@CC.UniVie.ac.at Jessica Yu jyy@merit.edu Romeo Zwart romeo@sara.nl