Why meshing is hot! A manifest to those who like to roam free
The Mesh - a vision of self-forming networks
This manifest attempts to present a vision of something that has
finally become possible – The Mesh. The Mesh, as we foresee it, is
an ecosystem of communicating entities – The Peers. The Peers populate,
use and constitute the nodes of the network, which forms The Mesh. Peers
are autonomous agents which are capable of individual, mobile action and communication. Information is communicated over self-forming, self-healing networks, which form a mesh topology. The Mesh consists of, and in fact is, its peers and the communication patterns they form.
The Meshed Peer
The Meshed Peer is a citizen of The Mesh. He has one or more devices
(e.g. phone, PDA, smart tag, intelligent tooth filling or piercing)
over which he communicates. Communication is physical, i.e. audiovisual, or digital, i.e. data. As the Meshed Peer is seen to be mobile, communication is in general assumed to be wireless. There is no network provider because “We are the Network!” and we, the Meshed Peers, provide the network by participating. While The Mesh is formed to satisfy the communication demands of the Meshed Peer, just as it is formed by precisely that communication,
it can actually provide for the Meshed Peer as well.
Due to the cellular nature of the network, peers can be shown either to
consume or produce bandwidth. Communities are free to use this information within whatever billing mechanism they choose. The Mesh provides peers with communication bandwidth and services on an ad-hoc basis. In theory, each node in The Mesh bills for any use of infrastructure it provides – while paying for any infrastructure it consumes. In practice, this is done by central control or monitoring of the routing mechanisms, which form the fabric of The Mesh.
Mesh vs. ad-hoc
Based on capabilities the IEEE 802.11 network protocol provides for ad-hoc
communication between two peers, the first commercially viable mesh systems have been developed recently. The first company to provide components and infrastructure to build mesh networks is Florida-based “MeshNetworks”. The difference between The Mesh, mesh topology, MeshNetworks’ technology and ad-hoc
communication is crucial inasmuch as it defines the borders between want-to-have and can-do. What this attempts to say, is that at this point we will qualify our statement that The Mesh can be built on available technology. The Mesh in capital letters is an abstract concept, which assumes and wishes to be nothing but a pure P2P network between nodes consenting on a basis of ad-hoc communication. In that form, we cannot hope to build it now nor
in the foreseeable future.
A mesh network as one would build now, would have to take some shortcuts,
such as seeding areas with Network Access Points (NAP). In order to avoid discussions of the nature “Was Napster a true P2P application or not?” we will use the term mesh network as a network consisting of optional wireless routers and network access points (NAP) in addition to the participating peers. These
infrastructure components are used to seed the mesh and make it function
even for the first (few) participant(s). Additionally - after and before all,
we are a market economy - there will be some form of formalized transaction management, authentification and accounting for each mesh. In practice and by nature of existing mechanisms, which we need to use, this means that centralized units will handle these supporting services.
Ad-hoc networks
Ad-hoc networks are sometimes referred to as infrastructure-less network architectures. The idea comes from research on how to equip mobile units such as rescue or search-and-destroy teams with communication. Participants could use handheld radios, each of which also acts as a relay for other nearby radios. The robust, self-healing properties of ad-hoc networks make them suitable for military use, either by mobile combatants, or to connect up smart dust sensors that would be sprayed across an area from an aircraft. Related to smart dust, which is used mainly in military research, is the concept of crowds
as a group of nodes that create their own network by spontaneously establishing local ad-hoc connections.
Using given properties of the IEEE 802.11 protocol, ad-hoc networks can be and have been realized using off-the-shelf technology. Assuming one has a suitable routing system, the main issue in pure ad-hoc networks is that of coverage. A basic requirement towards a network, which is to gain client acceptance, is that of coverage of a given area. For pure ad-hoc networks this translates into the problem of how many nodes a graph of ad-hoc connected nodes needs to have to cover that area. And then one needs to know the shape of that graph. An ad-hoc network without coverage of an area of reference should not be considered a mesh.
In practice, this constitutes the main problem of implementing commercially
viable ad-hoc networks. Unless a critical number of well-distributed peers
populate the network from the first moment, coverage cannot be guaranteed
and can therefore be considered as not given. We assume that this problem
is bound to find a solution in the not too far future. Once true pervasive,
ad-hoc networking becomes reality – The Mesh will be: an organism of communicating peers.
An organism of communicating peers
The Mesh as an ecosystem – and as opposed to the collection of autonomous
peers it contains – manifests the communication infrastructure between the peers. The nature of the mesh as an ecosystem will emerge as it forms and can be explored on an experimental and heuristic level at a later date. For now we just wish to emphasize some of the attributes, which emerge out of its organic nature:
The Mesh is based on the principle of a dynamic, living network forming
out of patterns of ad-hoc communication between peers. At the heart of the Mesh
lie the (routing and) forwarding mechanisms used.
Forwarding is designed so as to optimize throughput in face of ad-hoc
dis/connections while being economically rational
As an entity - an entity similar in nature to an ecosystem - the Mesh
has self-forming, self-modifying and reflective properties. It exists and
attempts to optimize its existence.
Instant reward possible
As opposed to many a Glasperlenspiel ("a-can-do-everything") introduced
under the guise of futurological or actual research, The Mesh can be built now.
We are fortunate enough to live at the precise crossroads in technical development, where all the individual building blocks of a new structure have emerged and are waiting to be woven together into the fabric of The Mesh.
A version appropriate to our limited understanding of what will be, can be assembled based on:
available P2P software
available WLAN/WiFi technology
available IP-telephony and VoIP frameworks
available Instant Messaging frameworks
available ad-hoc routing algorithms
Hic et nunc!
...sounds so much better than plain “Here and now!”, so please indulge the author. One of the goals of this manifest is to sound the call to start. In order to encourage research and work on The Mesh and confirm that the necessary
technologies and components really are available at this point in time and space, I will briefly define a few terms and then introduce the core technologies or components that could be used to build The Mesh – Here and Now!
Definitions
MESHED NETWORKS - Meshed networks refer to any number of sites arbitrarily
connected together with at least one loop. The meshed network can contain
spans at any permissible line rate. The sites within the meshed network
that can be reached from any other site by at least two distinct routes
form the mesh itself, while the remaining sites are spurs off of this mesh.
Meshed networks are often large rings with a number of subrings.
Virtually every network topology, except perhaps a pure linear span, can be
considered a meshed network. Obviously here the devil is in the detail,
or rather the addressing and routing mechanism used.
PERVASIVE COMMUNICATION - The network is the computer is the network....
Pervasive Communication can be thought of as:
cell-based
ubiquitous
location-aware
multi-channel or rather integrated communication
The Mesh becomes pervasive communication.
Meshnetworks alias Milcom
This company will be the Qualcomm of The Mesh – is what I say. MeshNetworks
is such an interesting company that they deserve their own section. If I’m not
completely mistaken, they will be the base technology provider for most
large-scale mesh networks in the near to mid-term future. See also
“Qualcomm business model” below.
MeshNetworks’s story starts at another company, Military Commercial (MILCOM) Technologies, which was founded to identify United States military (Department of Defense) technologies that are suitable for conversion to civilian adaptations. MILCOM Technologies founded MeshNetworks in February, 2000, to commercialize development work on mobile technology developed by the Aerospace/Communications Division of ITT Industries.
ITT has licensed this technology exclusively to MeshNetworks (and retained a
cross-license to incorporate MeshNetworks technology back into products for
military communications). Basically, the product now in the final stages of development is a mobile IP-based broadband wireless communications system capable of Voice Over IP (VoIP). Nodes/Peers using this technology dynamically form a network, constantly adding nearby nodes, dropping nodes that have moved, and keeping track of network routing. Such peers do not depend on “base nodes” to provide services to the network - “though external communication is obviously required for traffic destined for outside the battlefield”.
MeshNetworks plans to adopt the “Qualcomm business model”, i.e. manufacture
very little directly, such as chipsets and software (link layer, ad-hoc routing layer), and license its technology to numerous equipment manufacturers on a non-discriminatory basis. MeshNetworks expects the prices of its IP to be roughly comparable to those charged by Qualcomm and others.
ARACHNET - the Mesh according to MeshNetworks - The wireless WWW (W4). This is a fairly long-term project of MeshNetworks. Currently they focus on designing suitable boards and have produced a 3x5 inch circuit board with three chips. MeshNetworks expects to release software that can use 802.11b (and maybe 11a and 11g and possibly Bluetooth) to implement much (but not all) of the functionality of ArachNet. Longer term, MeshNetworks will develop its own RF chipsets that will, among other functions, incorporate highly accurate power management capabilities needed for full implementation of ArachNet technology. Ultimately an ArachNet node will be reduced to a PC Card and “embeddable” form factors. MeshNetworks expects nationwide deployment of ArachNet systems by 2005.
Technologies
If all these pages aimed for, were to pitch mesh technology, I would not quite
have mustered this enthusiasm. In fact what makes mesh technology so interesting, is that it provides the physical layer for interaction and communication on a much higher level. How high? Ideally, of course, the point of reference would be the individual person. However, given the heading of this paragraph this does not seem to be the appropriate level of abstraction.
This manifest claims that The Mesh can be built now, here and now, with available technology. This helps to restrict us in the consideration of technologies. The following sections present the most relevant clear and present technologies. Where needed, a short explanation of the connection to mesh technology is given.
MULTI-USER ROLE PLAYING GAMES (MU-RPG) - The term MU-RPG is used to describe role-playing or fighting games, where each user is represented by an avatar. This definition includes “shooting games” or hybrids such as Grand Theft Auto (GTA). The avatar personifies and instantiates the role the user is playing. The MU-RPG provides a world of reference, which the avatars roam. What makes MU-RPG avatars so interesting is that they can be molded to race, standing, abilities and attributes. Add to this the fact that modern MU-RPGs are highly structured and formalized software and communication products. The concept of avatar in this context then can be formalized to become a communicating software object, which can be used to create virtually antropomorhic peers of The Mesh.
PEER TO PEER (P2P) COMMUNICATION - As this chapter in itself could fill a book,
I shall take the shortcut to appeal to reader’s imagination. What is to be imagined is to be a good MIX of the properties and capabilities of systems such as Napster, Freenet, Gnutella in addition to state-of-the-art – leading but not bleeding edge - mobile voice and data communication. The result would combine at least
the file sharing and distributing capabilities of Freenet
mobile handheld VoIP terminals or PDAs as well as VoIP SW for PCs
wireless coverage in areas of reference
a virtual SIM module identifying the actual person
Based on this abstraction, I would like the reader to let his or her thoughts
roam free for a minute or two. In my case it resulted in the enthusiasm
that is permeated in these pages. Just think unified communication over
mesh networks including unified, instant messaging with voice and data and
a pgp-style web of trust between people that can be uniquely identified.
The type of applications people will ultimately used in mesh networks are
boundless but immensely powerful at the same time.
Wireless networks
The world is not a technical model. What this means, is that contrary to what
many self-professed geeks think, most of today’s issues regarding bandwidth,
its availability and its pricing have at their root not technical problems
but rather issues of ownership and billing. Taking this one step further
I even dare to postulate that current technology would allow us to
implement most of anything we need, if ownership and regulation would not matter. I will therefore approach the field of wireless networks not from a
technical perspective but rather on the basis of the following questions:
Who owns each mile? (Ether, last mile, backbone etc.)
Who are the subscribers or clients? (End user, corporate, public utility etc.)
Costing: how much does what cost whom?
Billing: who writes the bills? This leads to accounting and enforcement when the actual funds are transferred.
Based on these premises, wireless networks can be classified into the following four classes:
Satellite-provided bandwidth (for lack of more specific terminology)
Personal Radio Services, e.g. CB-Funk in Germany.
Telecommunication-operated wireless networks, e.g. GSM.
WLAN: wireless Ethernet based on the IEEE 802.11 family of standards.
Without going into any detailed technical analysis, it should be clear
that only Telco or WLAN operators can realistically provide the bandwidth
and infrastructure needed to build a real existierendes Mesh. Satellite-based
networks might fill a role as access points or gateways to fixed-line backbones. Personal Radio is a very low-bandwidth technology, which I would prefer not to use for the data-intensive services I would enjoy using. Personal Radio’s big advantage is that signals can travel quite far – in fact across the globe if need be. It might be useful to provide highly connected – and thus central – nodes with a reliable signaling mechanism. This could be used to coordinate highly referential mesh configuration data. One might envision a map of the central nodes and their states, encoded as data packets and (more or less) constantly broadcast between peers operating on Personal Radio frequencies.
For telco-operated networks we have as one of many issues, that these networks
do not allow truly independent mobile terminals. Even for packet-based communication which logically flows from one terminal to the next, traffic is first sent from terminal to operator, then travels the telco’s network infrastructure finally to be transferred from telco gateway to the addressed terminal. To a large extent this useless traffic, is based on different and sometimes even proprietary protocols. This, among many other issues, means that no true peer-to-peer communication can be established within existing cellular telco networks. This in turn means that mobile telco networks cannot realize many of the scalability advantages inherent in mesh networks.
All of this seems to point to WLAN technology to be the technology of choice,
when building The Mesh. It should be emphasized that, even if this translates
to the family of IEEE 802.11 protocols today, we do expect new protocols to emerge. It could also well be that such new protocols could be telco-operated.
Routing and forwarding
This is where the men separate from the boyz! Up to this point all technologies
or systems described exist, are known to be stable and are easily afforded.
However, there is one noticeable gap in the logical layers covered: we have suitable application protocols and a well-defined network up to the MAC-level (i.e. the world of IEEE 802.11). The Mesh however, is more than just a scattering of WLAN access points or hotspots. As stated before, the heart of The Mesh lies in its addressing, routing and tracing system.
This field is still wide-open to research.
It seems that the DARPA research programs, which ultimately spawned MeshNetworks, once again yielded quite beneficial results. It should be clear, however, that this is only the first industrially viable solution. New and better systems are to be gained for the price of some solid research and accompanying field projects. To give a sense of the beauty these routing systems will have, I quote from
“
A Performance Comparison of Multi-Hop Wireless Ad-Hoc Network Routing Protocols”:
“The actions taken by TORA can be described in terms of water flowing
downhill towards a destination node through a network of tubes that models
the routing state of the real network. The tubes represent links between nodes in the network, the junctions of tubes represent the nodes and the water in the tube represents the packets flowing towards their destination.”
This manifest calls out to all of those who like to roam free to participate
in building the routing system to do it.
