* [#goal Goal] ** [#members Members] ** [#mailing_list Mailing List] *** [#addressing Addressing] *** [#dns DNS] *** [#general_discussion General discussion] *** [#comments Comments] * [#archive_of_old_wgroutingaddressing_page Archive of old WGRoutingAddressing page] ! Goal To revisit the Routing, Addressing and DNS policies with a view to adding to, refining them, or just getting them implemented. !! Members We start again so join or rejoin. (The old members list is archived below) || GlennMcKechnie || NodeGES|| || DanFlett || NodeGMR|| || iViLe || NodeFUT|| || Exemplar || NodeFUU|| || || || || || || || || || || || || || || || !! Mailing List A new one is required. !!! Addressing Currently we are handling this via the node database by handing out /28 addresses, which is a block of 16 numbers. Do we ... * Continue as is ** Incorporate an online ipcalc tool (with instructions) to assist in selecting or rearranging allocated subnets * Modify the database to provide a choice of ... ** p2p subnets /30 - 2 hosts ** AP subnets /28 - 14 hosts, /27 - 30 hosts *** and explain why * Modifications would be aimed at leaving existing allocations as is? Whatever happened to the 172.x.x.x for routers? RIP? !!! DNS Nodes are starting to run [DomainNameSystem DNS] servers, what's our fqdn? So far we've had, or have... * nodexxx.mw.wifi * nodexxx.mwn * xxx.wireless.org.au * xxx.mwnode.com * and for more, see [DNSConvention] For various [DomainNameSystem DNS] servers -- DJBDNS - BIND9 - ... *Provide (via the wiki or plain text files?) generic DNS tables along with appropriate instructions How far do we go with DNS? If we NS everything then perhaps we use... * nodexxx.whatever for a nodes [AccessPoint] ('''AP''') * dhcpxxx.whatever for the APs '''clients''' * YYYxxx.whatever for [PointToPoint] links Status of ns.melbourne.org.au? !!! General discussion Reassess [Quagga] (zebra) It's currently working well in small clusters but appears to be failing when linking different regions. * Is it our area allocations (unlikely)? * implementation of them (probably)? * or our ad-hoc method of creating config files? Do we... * Supply generic configurations or practical examples for ospf.conf, zebra.conf, etc files. (These must include relevant comments to shorten the learning curve!) * Handball it to (work with) the coders to incoporate this into the IP allocation process generate a base config. ** or maybe as a seperate form so the process can be revisited. *** Input interfaces, IPaddress, netblock - output suitable config. *** also output link (yet to be created) to wiki notes, hints and tips (supplied by MelbWireless peoples experience on this network). Assess BGP Assess Frottle *Do we need it? *Is it more suitable for specific types of nodes ** GHO ** ILE Centralize linking to all wiki routing based material Add appropriate subnet division examples to wiki. !!! Comments ++++ ! Archive of old WGRoutingAddressing page Should we push all this stuff to the bottom of this page and start a new version of this page here? --DanFlett 2005/02/09 Yes -- done! -- [GlennMcKechnie] Goal To co-ordinate the Network Design Group. Members || Jason Brice || || || Dwayne |||| || Adrian Close |||| || Peter Rogers |||| || Matt Boyd ||[NodeBCG]|| || BrendanHiley ||[NodeDIH]%%%[NodeEAE]|| || Andrew Harcourt ||[NodeCHI]|| || TonyLangdon|| [NodeBCJ]|| || [Jeremy Lunn]|| [NodeDFC]|| || RogerVenning ||[NodeCIA] || || DavidArnold ||[NodeAAG] || || TristanGulyas ||[NodeFSA]|| || [Darius]|| [NodeFFB]|| || TimshelKnoll|| [NodeDDE] || || [AndrewGriffiths]|| || || Julien Goodwin ||[NodeGSO]%%% [NodeHUH] || || GlennMcKechnie || NodeGES|| || DanFlett || NodeGMR|| Mailing List (mailto:WGrouting@wireless.org.au WGrouting@wireless.org.au) %%%(mailto:Majordomo@wireless.org.au?body=subscribe%20wgrouting Subscribe) > Forget this mailing list address. A two months subscription returns 130+ A grade spam messages and not a useful post in sight.%%% Time for a change of list address. ---- Task List [ProblemStatement] and [Layer2Assumptions]. # Document current thinking on [AddressAllocation address allocation] # Produce a valid [OSPFArchitecture OSPF architecture] for the initial core network NOW, recognising that (a) the initial nodes will form the backbone / Area 0, and (b) we can afford to re-number / re-configure them while (c) non-area 0 nodes only come later - so we don't need to know how to set them up yet {/regexpicons/emoticons/emoticon-face7.png :-)} # Get working implementation including (a) ConfigsZebraOspfd , (b) gated configs? and (c) instructions on how to set this up from a default RedHat & [Debian] installs (& *BSD if anyone on this group runs it). # Scope [IPAllocation IP address allocation system requirements] and implement on-line. ## (These next two follow on from this but are not necessarily core group activities - see RouterNodeSetup) # ISOise a working [RouterNodeSetup] so that it will run/boot off CDROM being able also to boot of floppy and storing /etc /var and so forth on the floppy disk. (support of hard disks?) # Produce a [Debian] & [RedHat] package that will tack onto a default install to give configured, running routing nodes. # Determine a reasonable set of growth scenarios for Melbourne Wireless # Simulate routing table and address allocation under the prior scenarios in order to anticipate the behaviour of simplistic regional address assignment, 'carving' allocation, etc. Post simulation results # Come up with a [DNSConvention mesh best current practice] document that describes other non-mandatory aspects of participating in the wireless mesh. # Determine which multicast routing protocol to use, and implement multicast. # Devise [IPv6] addressing architecture and provide [IPv6] implementation details] as well as [6BONE] Peering. Process for acquiring addresses This is now active via the node database. Handing out /28 addresses, which is a block of 16 numbers. Comments Does anyone have the skills to port mobile mesh to windows? RogerVenning does... but would need convincing before embarking on such a project. -- You would be a legend, adored globally by giving all wireless groups a standard routing platform. How's that? {/regexpicons/emoticons/emoticon-face3.png ;)} ([RogerVenning]: send me an email and we'll talk about it) I think we really need to organize a few meetings and/or monthy/bi-monthly meetings where people actually turn up to learn about Zebra, how OSPF is going to be implemented, discuss IPv6 (which is very very kool), and other things as they pop up. --BrendanHiley Sounds like a great idea Brendan. Another possibility, now that the meetings are quite large, is to find a venue where several things can happen at once. But I spose thats something WGMeetings can work on... --JamesHealy Can someone please think of a more suitable "Goal" for this group? "To co-ordinate the Network Design Group" gives no useful information... --TysonClugg * How about to create a cross platform no brainer install routing protocol with automatic link discovery and route optimisation. nothing difficult or anything {/regexpicons/emoticons/emoticon-face3.png ;)} --[Drew] I will be giving a brief summary at the [Meeting20020712 July 2002 meeting] of the Aarnet IPv6 workshop which I attended on 24-25 June 2002. I will make myself available for further assistance to this working group for any IPv6 related issues. --TysonClugg ++++