Vladimir Radunovic   09 Sep 2013   Internet Governance, Diplo Blog

Printer Friendly and PDF

Back to Part 1: so - could we then simply boost the mesh networks and connect all of them in a similar mesh way to get the globally green Internet 2?

Probably not!

Killing dreams may be better than dreaming awake. Let’s start with performance.

The real bandwidth within mesh networks might not be very comforting (and certainly a dozen times less than the 150Mbps which is a theoretical maximum at several metres’ distance only). This especially goes for networks in densely populated areas where free wireless spectrum used can be highly saturated. Additionally, current protocols for routing the data in mesh networks can cause significant latency (delays in delivery): the more ‘hops’ (users in between sender and receiver that relay the data), the greater the latency is. While simple services may work fairly well even in big mesh networks, real-time applications like Voice over IP (VoIP) or video streaming are still very challenging within networks of a few thousand users.

Moving on to myth-busting the five alleged features of the community mesh networks listed above:

  1. Surveillance: Mesh networks need to use wireless for practical reasons. Wireless networks are very easy to eavesdrop on: while for cables one needs to physically plug in to the network, wireless spectrum can be ‘sniffed’ imperceptibly from a safe distance with, if needed, by breaking the (usually weak) encryption. The lack of central connection points makes the task more complicated, true – and the NSA would probably not bother sending agents everywhere; yet many other (far less skilled and equipped) folks might.
  2. Kill-switch and censorship: True, an authoritarian regime would hardly be able to interfere with the internal traffic of mesh networks, even though it would still be easy to censor the incoming Internet traffic at the entry points into mesh networks. Unfortunately, a determined sovereign could easily activate the kill-switch for wireless networks of any type: deploying a few high-powered transmitters around the city to emit noise signal would jam all the wireless communications on a certain frequency.
  3. Net neutrality: Due to its decentralised structure, there is no way to manage the traffic within the entire mesh network without the consent of all users. Yet, the users themselves could possibly deploy software to manage the relaying traffic (passing through their own computers from one user to another), thereby saving their own bandwidth occasionally while throttling the data flow of some others. This may not require high skills (such software is available), and might in fact appear realistic in bigger networks where congestion occurs.
  4. Last mile access costs: True, the cost of access to a mesh network is a one-off for  wireless equipment, which is symbolic. If the entire mesh network is at some point connected to the Internet, subscription costs may be equally shared by all users, or even according to the actual data flow – as agreed internally.
  5. Intellectual property rights: The centralised control of file sharing is not applicable. Yet, since the geographical coverage of a network is very limited, with identifiable inhabitants and under a unique jurisdiction, the incentivised authorities might sniff the wireless traffic, identify copyright breaches, and possibly prosecute the offenders.

Dreams of a global mesh

Even though not ideal, mesh network do reduce dependence on big industry and vulnerability to mighty foreign governments. Let us not forget, however, that – in spite of the possible abundance of customised local services within individual mesh networks – the end-users would still likely be inclined to use the global Internet heavily in order to reach out to global content and distant friends. Could the same mesh principle then be applied to connect smaller mesh networks among themselves – forming a mesh network of mesh networks, i.e. Internet2?

Not really. Such an interconnection would require high bandwidth backbone connections between meshes – a bandwidth demand that could be provided only by fibre. Mesh networks connected by fibre instead of wireless would be way too impractical and expensive. Even if wireless technology advanced to the extent that it could compete with fibre (which is not likely), connecting mesh networks globally would be tricky. In one scenario, thousands or millions of orbiting Google-like balloons would be needed, which is costly and hard to maintain so only big industry could run such an endeavour (and then we’re back to centralised corporate dominance). The other scenario sees global utilisation of ‘whitespace’ frequencies (television frequencies freed after the digitalisation process which have a metropolitan area outreach and will increasingly be used for wireless communications); however, governments are the ones to decide on the use of this spectrum, the related technology for end-users is still not widespread, and – most importantly – these windows of wireless spectrum are also scarce resources and would be depleted by a massive demand, leading to congestion and dysfunctionality.

Cloudlets can announce better weather

By no means are community mesh networks bad, of course – they are more than affordable, useful, and inspiring, especially in developing countries and to connect distant areas. But they are not ‘a silver bullet’ for IG challenges.

If applied to software and services, however, the mesh concept may further reduce end-user dependence on big industry and mighty foreign governments – i.e. on giant clouds behind major global services (like Facebook, Google, or Amazon). Imagine a distributed or a mesh social network (like the Diaspora project) which would store personal data in your own storage or on small ‘cloudlets’ within your community network, and share them with friends only, directly through a peer-to-peer connection.

Current global clouds resemble giant, dark, stormy cumulonimbus. Up in the sky, it is cirrocumulus – a patch of small white puffy cloudlets – that often forecasts improving weather after the rain. Digital cloudlets may do the same for an open Internet.

 

Leave a Reply

Plain text

  • No HTML tags allowed.
  • Lines and paragraphs break automatically.
  • Web page addresses and email addresses turn into links automatically.
  • You may use [view:name=display=args] tags to display views.
CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.