created 25.07.2000 last ammended 3.08.2000 author Julian Priest last edited by James Stevens 03.09.2000 consume the net #aims #orgs #folk #nodes #services #slogans #peering #benefits #technologies #finance #gui #core kit #tests #legal #mailing list #aims The aims of consume.net are; define a sustainable network development utilise low cost and licence free radio IP systems optimise infrastructual expenditure and reduce network connectivity costs increase network speed re-distribute access promote common ownership increase resilliance aggregate avaliable 'internet' bandwidth develop top level peering status and of course have more fun model 1 The initial working pilot makes certain assumptions. This network will be used to test these assumptions and others to arrive at a sustainable network model. 1. assume persistent connections at each node 2. there is spare capacity at each connection 3. this spare capacity can be offered to the netwrok 4. all interconnects are 2 way 5. the owner of interconnects benefits to the extent of their total interconnect bandwidth 6. the minimum speed available to a connection owner is the speed of the line they own. 7. it is in the interest of a node owner to interconnect as fast as possible 8. it is in the benefit of the entire network for nodes to be added. #orgs These are organisations that have either confirmed to be node [~] signalled an interest in the project [*] or are contacts of project members who will be invited to get involved in the near future. If you are not listed and should be or are unhappy about your inclusion at this stage then please contact james@spc.org. mediumrare~ spc.org* lateral lux~ thought interorum foundry~ ellipsis romandson darq planet hackney yesmate~ lovely tecc elektrosmog LShift A.L. Digital Ltd The Bunker static yootv comala planetme singlelton silver server ProteinTV~ #folk A list of people who are engaged in the project or have requested information. julian@mediumrare.net james@spc.org tim@red56.co.uk sacha@thought.co.uk ian@darq.net vaughan@psyanide.net richard@binaryrefinery.com (richard jones) angus@planethackney.com (angus dunwoody-dunn) alexeiblinov@hotmail.com (alexei blinov) justin@raylab.com paul@yesmate.com honor@va.com.au col@clinkstreet.com lee@lateral.net kevin@frontier.com andrew@lovely.net john@lovely.net howard@ihac.org stephan@comala.net julian@planetme.com susan@singlelaw.com oscar@sil.at christian@mediumrare.net #nodes initial locations for transmitters are buildings which have power and access to high points such as a roof or tower. Proximity to another node and permenant/regular internet connections are essential. The initial locations will be experimental and so it is envisaged that most of them will have technically minded people who are willing to contribute. #existing clink street studios se1 foundry e1 ellipsis e1 shoreditch town hall e1 lateral (charlotte road) e1 lux building (hoxton square) e1 percy street wc2 #probable Truman Bricklane Interface - Commercial street #possible/interesting lwt building itn building *centrepoint southwark cathedral se1 #services the network is designed to facilitate some higher bandwidth activities and part of its remit is to provide a space for members to experiment with real world high speed or broadband applications without the prohibitive costs that are associated with those activities in more bandwidth limited spheres. The ability of the network to attract services is also a key to its success in publishing to the wider network and an aggregation of content on the network will make peering with it attractive to other networks. In return for access to the content and high speed interconnect the wider network providers may peer.. eg reduce the connect fee from its current high levels to zero. streaming media internet access game servers public programming (frequency clock sheduling) #slogans personal telco pico peering #peering peering normally takes place between large network providers to mediate traffic into and out off their networks. The peering model of consume.net will create peers between individual unaggregated network nodes. The peering agreement will allow access to the network on the condition that its terms are met by the interconnecting network. The peering agreement passes to its next neighbour. In the instance of wireless networks the peering is also proximity based. The consume.net peering agreement may have 2 parts. The first will specify how nodes within the network peer with each other and the second specifies how the netowrk peers with other networks. Ideally these would be present in the same general agreement. Legal work is needed to see how this agreement impacts on existing network contracts. #benefits access to content agregation of spare bandwidth local high speed networking agregated bargaining power short circuiting the local loop (reducing the cost of local loop connection) #technologies #wireless ethernet Wireless ethernet conforming to the 802.11 standards is a license free networking technology which runs at up to 8 mbits. It can be thought of as the networking equivalent of CB radio. The technology is low cost and standards based and comes actually plays to some of ethernets stronger points (ethernet was origninally developed as a wireless broadcast medium). There are 2 license free bandwidths that may be used in the uk 2.4 Ghz and 3.5 Ghz. Most activity is in the 2.4 Ghz range which leaves a virtually unused band which consume.net will exploit. There is much activity in the states in the wireless isp industry and as a result most of the common problems and there solutions are in the public domain. The equipment is also fairly well supported for drivers under linux and bsd. The low cost of the medium makes it ideal for consume's initial phase. It may be used in directional or omni modes depending on arial design. Directional instalations obviously fit more bandwdith into a smaller physical space and are prefered. Omnis are useful for street wide coverage etc. It may be that the backbone of the netywork is initially done over 3.5ghz directional and the last hop to laptops etc. is done over 2.4 ghz gear. There are many vendors such as lucent, breezenet, nwn and many more. The commodity status of the equipment means it is readily available and prices are headed only one way .. down. There are also developments and additions to the standard that will increase the bandwidth with some vendors suggesting a 24 Mbit version and the hiperlan 2 consritium proposing 56 MBit versions. #router each node has a router which routes (der) and serves as a container for the mediating software which is router based. It should also contain a small web server which publicises the legal frame work of the network and the conditions for usage (npl). Initially it is proposed that this router be based on a floopy based installation of one of the free unix os's. A selection of these are trinux, pico bsd and linuxrouter.org. The PC that serves as host should have a minimum spec of a PC (486 +) 16MB Ram + a floppy disk drive. 2 PCI cards. One PCi Card is for ethernet access the other to hold the wireless card. This enables us to have a cheap flexible platform for developing the network on, and a software platform that lets us be flexible about how the supporting software is implemented. The router also should manage network traffic so that allocation is controlled and the user always retains as much bandwidth as they are donating to the network. #free space laser This is a new wireless eyesafe high speed networking technology. Data is transmitted over point to point infra red laser at very high speeds (622Mbit and 1 Gbit) Older versions of this were very expensive and prone to misallignements, pigeons etc. New technologies mount several autosensing units which train themseleves on neighbouring installations to form a high speed mesh of connections at very high speeds. We are hoping to gain funding to install a metropolitan network of these devices to form a high speed backbone. Some vendors are airfiber.com (distributed by nortel) and terabeam.com #lans and other networks Donnate your LAN or other fixed network to consume.net. Bandwidth doesn't have to be wireless. A slong as it provides a route from one place to another and expands the network then it is valuable. A piece of category 5 through a wall is a cheaper bandwith solution than a wireless terminal. Agregating a building means that you increase your bandwidth and reduce your costs. #ipv6 Give every device on the network its own address. ipv6 also has advantages in terms of security and mobility. The utilisation of ipv6 will allow the network to have a fair chance at populating the new address space and give the network a technological superiorty to the existing network. It will also encourage other networks to join the network as a possible upgrade path from ip4 to ip6. try theseORiNOCO #finance #node based (individual node financing) the core of the network is its users nodes. These finance their own nodes thus making their local bandwidth increase and adding resources to the network. #sponsored (games companies initially) We have some interest from games comapnies who wish to be able to give their players access to a high speed network as a marketing exercsie fro their games #service funded It is in the interests of service providers who wish to sell services or bandwidth to people on the network to increase the speed of access to them through the network. This is done by them making infrastructure investments into the network surrounding them. Initially these may be isps (band x) or other network services such as satellite providers (gloebcast) etc. #public bodies funded There has been some interest from councils who see this as a means of providing low cost high speed access to people in their areas. Possible councils are haringey and hoxton. Councils themselves have access to government and european funding to develop public access technology initiatives and this could fit the bill nicely ffor them as it is inherently a publicly held enterprise #gui There is an associated project to map the network and edge towrards a new interface paradigm for network based broadband applications. This toolkit will be used as a common interface to services that exploit the potential of the netwrok. Initial ideas suggest a gl based stalker like network mapping tool (similar to iods "into the web"/stalker) crossed with a media interface. #core kit pcmcia card -station adapters pci card (converter) access points -£650 antennas -£150 installation 3 amp power supply pci based computer 2 slots 16mb ram floppy disk ( no hd!) ethernet switch and category 5 cabling rooftop or near line of sight to nearby locations #tests use three modems and three routers to build a network with network connections to isps and a 10baset network to simulate a network faster than the connections. #legal develop a network public license that creates a single node personal isp. This will set the rules by which nodes connect. Nodes that use services that lie under the npl will mediate their interconnection through the npl. It is enforced by the routing software. #mailing list
If you wish to be included in the pilot phase (now) or want to know more please....subscribe review the archive here