The Imminent Rise of Social Cloud Computing

Cloud computing is the distribution of services like information storage and processing as if they were utilities, rather like electricity sold over a power grid. The idea is that people use and pay for only as much as they need.

But one problem with this model is trust. How can users be sure that all the nodes in the cloud can be trusted, that there aren’t some nodes that are malicious? Such nodes might disrupt information processing tasks by refusing to co-operate or by sending back false data, for instance.

Today, Abedelaziz Mohaisen and pals at the University of Minnesota in Minneapolis suggest a fix. They say one solution is to distribute tasks only to people you trust, as defined by your social network.

Their idea is to ask members of a social network to make their computers available for processing tasks when the computers are not otherwise being used, rather in the manner of distributed computing programs such as SETI@Home. Then whenever a user outsources a task, the burden is passed to their friends; in other words, to nodes directly connected to them on their social network.

In essence, Mohaisen and co are saying that if you share your photos, videos and life events with other people on a social network, you will also trust them to store and process information for you too. That seems reasonable.

But how would such a system work? These guys have also modelled the way such a distributed computing system would schedule, distribute and complete tasks in several real-world networks. One interesting result is that social networks seem to be well matched to the task in hand. These networks ensure that tasks are distributed in way that most of them can be completed relatively quickly, provided that fewer than 30 per cent of users are outsourcing at any one time.

Whether that will turn out to be the case in a real network, we’ll have to wait and see.

If this idea takes off, it won’t be the first time that social networks have been used as a backbone for other tasks. Examples include information sharing, anonymous communications and so-called Sybil defences in which malicious users can be spotted on the basis that they are likely to have few friends.

So it seems reasonable to imagine that a distributed computing service could piggyback on an existing social network too and that a company like Facebook might consider such a business model with interest.

Ref: arxiv.org/abs/1112.2254: SocialCloud: Using Social Networks for Building Distributed Computing Services

Keep Reading

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.