Extending The Internet: The Peernet OpenMesh
Extending Internet reach and accessibility while making it more resilient, and immune to possible controls and manipulations has been one of the major concerns addressed, and although some significant steps forward, at least in terms of awareness, have been made in this direction, bridging the digital divide and safeguarding Internet net neutrality has remained for now only a purely economic equation.
Photo credit: Circotasu
Individuals, like you and me have for now very few tools to influence the laws and regulations that determine how and where new cables or new Internet access technologies are deployed and what ISPs, telcos and our own government decide with our votes and money.
But we may possibly be at a turning point.
For the first time individuals can start affecting the way the future infrastructure of the Internet will develop. For the first time we have the tools and know-how to start thinking and experimenting with new ways to extend the Internet beyond the factual limits established by the present telecoms and corporation-controlled infrastructure.
The idea is one of extending the Internet beyond its present physical reach. If more people could extend and relay via the use of new and hybrid technologies their own unused bandwidth to other people not just a few feet away, enormous benefits could be provided to all those lacking quality high-speed Internet connections (let alone any connection at all) in cities and rural areas just about everywhere.
It is not that that those providing access from telcos to ISPs are bad or trying to do anything nasty, but if it comes to something as vital and fundamental as our own planetary communication infrastructure we better evaluate well whether, when the technological conditions exist, it wouldn't be better if each one of us owned and powered its direct node to the Internet, or redistributed to a much higher degree what he did not use, rather than having to buy exclusively into an existing established commercial infrastructure and depend fully on it (and on whom has the power to control it, limit it, censor it, or close it)?
It would also be good to consider whether it would be good to collectively safeguard the contents and services we use and like most on the Web and made them fully resilient and out of the control of any one entity out there.
With trans-oceanic submarine telecommunication cables being cut on the seafloor of our planet with no-one apparently being able to stop or even explain the nature of such rehearsals, or by looking at a recent court action resulting in WikiLeaks being taken off line, it appears increasingly likely that access to the Internet may not be a certainty we can count on in the foreseeable future.
In this light, it would appear wise to study and analyze the opportunity to backup the whole Internet including its content and key services, so that in case a natural or an artificial disaster brings the official Internet down, we may have a fully redundant alternate system through which we can continue our planetary conversation.
While the basic architecture of the net does protect to some degree against these dangers, Sepp Hasslberger brings forward the idea that we might profit from developing a way to "back up the internet" so that, even if there are major disruptions, we still have a workable means of communication, data storage and exchange of ideas.
"My dream is a p2p application that uses some of the free hard disk storage space on our personal computers to redundantly back up the net and allow work to continue more or less seamlessly in the event of a major catastrophe.
Would such a thing be doable? What do you think?
Is anyone already working on this?"
Cyber Warfare on the Rise - Don't rely on the Internet
Alex Ansary - Cyber Warfare on the Rise - Don't rely on the internet - a video I run into yesterday which portrays some of the concern expressed in part in this article
What follows is an edited, revised and extended version of what originally written across a series of spontaneous posts by Sepp Hasslberger on the P2P community forum. I have re-organized this content, edited much of it, and added my own contributions after having spoken with the author wherever I felt I could add more clarity and depth to his original idea.
This is just a first draft of an idea that wants to develop more. There is no definite recipe or solution yet available but simply a desire by a few like-minded people to take this further and see where it can take them.
The Medium Is The Mess
While the original internet infrastructure was one of peer connections between university mainframes, the emergence of the world wide web has led to a paradigm shift, as recently argued by Simon Edhouse in The Medium is the Mess.
Today we have a dominant server-client architecture which we inherited from the world wide web that was established in the 1980s. We have clients and servers, and in physical infrastructure, we have consumers and access providers.
The paradigm is great for business, but it does not place everyone on equal grounds. The issues of net neutrality and access censorship have already provided some good indications of what the larger implications of this may be.
The Web and the entire Internet structure are corporation-controlled on which we are mere guests, much like the first people sending email and discussing on the Usenet, timidly using some of the bandwidth that was there for entirely different reasons.
So, what keeps us from overcoming this legacy architecture with all of its drawbacks, is that the overwhelming majority of transactions today is firmly grounded in the client/ server paradigm.
The worthwhile activity we should engage in is one of shining a big light on the inequities of the currently dominant client-server paradigm, and one of building increased consumer awareness of the client/server paradigm flaws so that large groups of consumers, (and the people who influence them) are more inclined to break away from this web-lock-in, and can experiment with the possibilities of possible new alternative systems.
What is needed, it seems, is a strategy towards the transformation of the web-dominated Internet of today into a possible 'Peernet' of tomorrow.
(The term Peernet has not been coined by Sepp Hasslberger and it has been used before, namely in a paper entitled "PeerNet: Pushing Peer-to-Peer Down the Stack" and addressing how a network with separation between address and identity would be designed. This would be a peer-to-peer-based network layer for large networks. Note that such PeerNet is not an overlay on top of IP, it is an alternative to the IP layer.)
Peernet would be an ideal extension of the present Web, but capable of running on a fully parallel, fully P2P-based infrastructure. Peernet would not need centralized servers and not everyone would have to go through an Internet Service Provider to get on the net.
Everyone would be a relay node capable of receiving and sending data to everyone else, while data and services would reside in a "cloud" made possible dynamically by all those connected to this virtual network at any one time.
There are many ways to envision this and at this stage they are all worthwhile consideration. Some see the possible framework made up by a relatively few traditional users hooking up to the mainstream Internet and sharing back access to the majority. Others prefer to view a gradual progress to a fully parallel and independent P2P-based net-like infrastructure. There are obvious and apparently unsurmountable problems on all sides, but notwithstanding these technology is bringing some interesting and disruptive new technologies to the open market while some quite interesting ideas have already been floated in the P2P community forum where the discussion is developing.
The difference between today's Web and a hypothetical future Peernet would be as great as that between the traditional media (both print and TV) and the World Wide Web.
By empowering individuals to take back full control of a global peer-to-peer communication infrastructure there would be hope of development of alternative economic and commercial realities which have no fertile soil on which to grow right now.
Eventually, for Peernet to be an idea empowering everyone, it needs to be based on its own P2P infrastructure quite independent of the hardware and even the connectivity that powers today's internet.
Digital Divide and Connectivity
There is also a persistent problem with bringing connectivity, especially broadband, to consumers.
Bridging the last mile is not easy. Cities attempting to provide that access with city-wide mesh networks are having trouble with their plans. One recent example: Long Island Wireless, Short. Portland Oregon is another trouble spot, where connectivity has been contracted and promised but is not really materializing. Other news and reports seem to indicate that mobile broadband is "spreading like wildfire", but the commercial interests involved make it for now unclear whether this will provide a true wide extension of access to the Internet for many, or if it will be just a rich market for the few that can afford it inside large populated metropolitan areas.
A simpler way to bridge that divide between cable and consumers would perhaps be a peer-to-peer consumer built and consumer-maintained mesh network that could be linked to the main, traditional Internet pipes through comparatively few access points.
Perhaps that infrastructure could be improved by adding a layer of mobile connectivity through opening mobile devices up to direct connection between each other, rather than routing all traffic exclusively through the providers.
It would be economically efficient not only in terms of investment but also in terms of traffic.
P4P software as recently tested by Verizon could keep much of the file sharing traffic on the "local" mesh.
There Is Not Just The Web: Other "Nets" Can Be Built Over and Beyond The Present Internet
Right now the Internet provides us with a basic communication infrastructure which consists of the links between all computers making up the internet. The links between nodes are provided by telecom lines, large communication backbones and access providers.
On top of this infrastructure sits a layer of communications protocols we call today the Web.
The point we need to start seeing is that we should not confuse the Web with the Internet by using these terms in a loosely interchangeable way.
Other communication platforms can also sit on the Internet's connectivity protocols.
A good example is Skype. It is a P2P application, and when people use it, they are not actually operating on the Web in the traditional way, but next to it.
Separate from both the Net and the Internet as defined by Edhouse, peer-to-peer could give rise to a different type of entity that links us up without being subject to controls except those we may wish to allow.
As a matter of fact P2P may in time give rise to something entirely different from both the Internet and the Web as we are conceiving them now.
What Does a Peernet Look Like?
A popular peer-to-peer based network would be best situated on a physical infrastructure built in accordance with P2P principles.
The existing core Internet backbone, (powered incidentally by 340 separate Internet backbone providers globally) while subject to degrees of segmented private ownership is still largely open, flexible, and is a medium that does not overly discriminate against innovation, even in the P2P field.
One just has to reflect on Skype's fairly successful disruptive activity in the telco arena (246M unique downloads, and 27M regular users) to see clear evidence of the power of a P2P system to take on the status quo and kick butt.
It is fascinating to envision real peer-to-peer connectivity starting with consumer-driven mesh networks based on WiFi or WiMax or a combination of both, and a gradual separation from today's internet even for long range connectivity, which could in a first instance could be driven by P2P radio bridges. Mobile device mesh networks could be part of this.
As almost everyone has a mobile phone today, it would not be to alien to consider the development of stacks like Android that would run on mobile phones and which would allow for them to become not just receiving units but full relay stations.
At present there would be many technical and infrastructural obstacles to realize this but technology appears to be fast cracking solutions for each one of them.
Backing Up The Internet
At the same time as we develop a new P2P communication infrastructure parallel to the existing Internet, we might also get so ambitious to think of developing a system that could help us back up the whole internet (or what would appear as such) outside the Internet, and possibly through this very Peernet framework.
The idea is the one of dynamically backing up Internet data on a cloud of interconnected computers, possibly with a novel way of distributed and redundant data storage inspired by an algorithm that mimics holographic storage of data. Sounds too far fetched?
There is a huge potential in personal computer hard disks, optical and 3D storage and new breakthroughs are happening constantly. The hope that one day you will be able to record and store all-of-the-data-out-there is already become a near-certainty.
Remember also that in a very-large sized network as the one we are envisioning, there would be a huge quantity of computational power available for reconstructing that data residing everywhere and nowhere, on the cloud of networked computers at any given time.
Not to mention that, if we just wanted, communications and digital identities could be re-invented in a secure and spam-free manner.
Eventually, this P2P net (the Peernet) could grow so pervasive that we would allow it to take over most of the functions of today's Internet while adding new things we never dreamed were possible.
How To Build The Open Mesh
But the question remains: How to motivate consumers to start building this open mesh?
Here is an interesting idea: in The Open Mesh Revolution, Daily Wireless reports about the launch of Open-Mesh, founded by Michael Burmeister-Brown, developer of the Dashboard Software that made managing dozens, even hundreds, of Meraki wi-fi repeaters fast and easy to use.
Here some of the key specs for these truly marvelous units:
- Inexpensive. Open-Mesh WiFi repeaters cost $49 each or $39.95 (qty 20)
- Ad free. Open-Mesh promises they will never push ads into your networks. You decide what, if any, content you want to display.
- 100% open source and deployed on top of OpenWRT. You can change anything. Open Mesh is open source and promises to stay that way, unlike Meraki and its Spain-based competitor FON.
- Re-flashable firmware.
- Free administration, alerting and mapping via the Dashboard management system which allows you to configure the ESSID, splash page, passwords, and bandwidth allocation of your networks.
- Auto-configuration. Easy creation of a neighborhood or apartment network. Supports other management systems as well.
While the reach of the routers suggested by Open Mesh isn't that great with only about a 100 ft range, this new technology is certainly a significant step towards making the consumer-based open mesh net work become a reality.
Not to be missed in this story is also ROBIN (ROuting Batman Inside) the Italian Open Source mesh network project, developed by Antonio Anselmi and deployed on top of OpenWRT kamikaze, running on the new Open Mesh routers as well as on any Atheros AP51 routers such as Meraki Mini or La Fonera and using the BATMAN routing algorithm (www.open-mesh-net/batman).
ROBIN's mission is to produce an Open WiFi Mesh Platform optimized for low-income,community wireless, education and the developing world. Built completely upon open source software projects, it will spread a wired internet connection such as a DSL throughout an apartment complex, neighborhood, village or school, and work on a variety of commonly available, low-cost hardware.
ROBIN is a zero-configuration appliance (plug & play solution): all you need is almost one DSL upstream DHCP capable router offering Internet connectivity where you connect a node (acting as gateway node), others nodes (client repeaters) have only to be powered on.
Presently ROBIN is open for testing and evaluation by simple users, community wireless, commercial and non-profit organizations. Among these, NetEquality has developed own ROBIN backend server for remote management of ROBIN mesh clouds (dashboard, update, upgrade, statistics and graphs, ...) also offering dedicated pre-flashed routers. The backend server is part of the NetEquality project open-mesh.com
Would 'Peernet' Be Desirable?
What sparked the idea of Peernet was the realization that we should have a 'plan B' for the internet. So in case of a major catastrophe, we would not lose connectivity and the ability to maintain communications directly with each other.
Our interconnectivity today depends mostly on physical connections such as optical cables which have recently proven to be very vulnerable and which may go down in any major catastrophic event.
Also the mainframe computers on which we depend to act as servers are not immune either to such situations.
A distributed architecture, based on open-standards, that can re-construct its data and which can function regardless of the number of peers involved, seems ideal for guarding against such possible future catastrophic changes.
But not only that.
With the experience we have gained from the Internet, Peernet could be designed to be spam-free and secure, and impervious to any outside interference.
It could also provide a fertile ground for establishing new and effective extensions of our present monetary systems (see also this open money discussion community for more references on this topic) and might have other advantages that are not yet as obvious.
Potential Killer Applications on The Peernet
There are a number of existing and possible applications that could drive high up the perceived immediate value of such a P2P-based parallel net.
Identifying specific killer applications that could drive the initial development and aggregation of multiple Peernet-interested parties would be particularly useful at this stage.
Here a few ideas to start with:
Communication Tools. A company called Mermaid has just come out with a full set of P2P applications that allow the full set of direct communications in a P2P network of any size. News broadcasting, audio and video communication, screen-sharing, movie-casting and a lot more. The initial suite is already quite impressive and nonetheless the UI is yet very simple and somewhat rudimentary, the apps already work and the company is promising versions of all its P2P tools for Mac and Linux as well.
Identity. Much work has been done on identification, and in resolving the issues of security, privacy, spam and identity on the Internet. In the presence of a new P2P based global network new digital identification systems centered around individual users and not around services on which register and define our identities could be more easily introduced. This could potentially provide a level of security unavailable and undreamed of on the Internet today. This would require lots of complementary (and difficult) steps, but nonetheless the path to provide a truly reliable digital identification system. For more info and inspiration in this direction please Wes Kussmaul white paper: "Have Identities Before You Manage Them" (PDF) or better yet the ID-Commons.
Alternative currencies and payment systems. Existing payment systems could be easily extended to the Peernet while new alternative payments system could also emerge with less resistance. Money could flow freely on such a network and it could be quite different from what we consider money today. For more hints in this direction please check out Ryan Fugger's Ripple Project.
I am sure there a lot of other possibilities, and I openly invite you to contribute them either here on the Comments area below or on the Peernet forum discussion on Ning.
Related Resources, Info and Tools
Amateur groups building up such ad hoc multi hop networks:
- http://freifunk.net in Germany
- http://reseaucitoyen.be in Belgium and Northern France
- some in the country side , such as one covering a whole danish region http://djurslands.net/ and
- and even one in the Himalayas / Dharamsala
- An article in Technology Review about Android (Android Calling), a standard for cellular devices promoted by Google, may bring us a good step closer.
- Also check out Martien van Steenbergen's comment to another string to the "Backing up the internet in a P2P 'cloud'" post. He links a post outlining the basic protocol for a software that could keep us all linked up properly. Very interesting approach.
- The Germans are great on linking up with radio hardware running IP protocols. The network seems to be growing already.
- Open Moko
- Martien van Steenbergen - Martien's Armillaria project
- The Medium is the Mess
- Backing up the internet in a P2P 'cloud'
- Could 'Peernet' be separate from today's internet infrastructure?
Martien van Steenbergen
Martien van Steenbergen wrote up a conceptual protocol that exactly does what you mention: storing redundant copies of every bit of information fully distributed over the net. I called it the Wizard, Rabbit and Treasurer. The Wizard is the illusionist, serving you your information in a smart way so you have the perception that it's always there. Anytime, anywhere, any device. The Rabbit is the one locally serving the Wizard and other, remote, Rabbits. In fact, the Rabbits implement the peer-to-peer infrastructure for volatile data. The Rabbit uses a local Wizard to permanently store data on non-volatile media.
For more info on this conceptual protocol please see http://wiki.aardrock.com/Wizard_Rabbit_Treasurer
Martien van Steenbergen also commented that Sun's ZFS contains many pearls for implementing what is describe here. "Simply connect the storage devices through a network rather than through USB, SCSI or IDE. Must fix the timing (latency) issues of course" he writes.
Pre-compiled images download (for Meraki Mini and La Fonera)
Easy-flash tool (only for La Fonera, with beta rv1113 firmware preloaded)
Originally written by Robin Good for Master New Media with significant contributions and original content by Sepp Hasslberger and the P2P Foundation blog. First published as "Extending The Internet: The Peernet OpenMesh" on March 19th 2008.
blog comments powered by Disqus