Xavi Masip, Scientific Director (UPC)
Sharing, sharing, sharing….and more sharing, it seems we all, as a society are being pushed towards a world where sharing becomes a must. However, questions such as what does sharing mean or what the real implications of sharing may be, are remaining issues yet demanding additional efforts from the whole scientific community.
Obviously, from a lexical perspective, its meaning is pretty clear, but the way this well accepted and widely used concept is applied to technology still deserves some specific attention. Indeed, we all heard about innovative services leveraging the sharing concept in different verticals, such as car sharing or collaborative housing. But where are these services inferred from? In fact, sharing as a concept applied to technology is not a disruptive idea, and many distinct successfully deployed examples can be found in the networking and computing fields, such as volunteer computing, community networks or the real PlanetLab platform.
|Volunteer computing formally started in 1995 with the first volunteer computing project, called GIMPS (Great Internet Mersenne Prime Search, http://www.mersenne.org/) and is defined as a type of distributed computing supported by computing resources donated by the own computer owners to cover research projects needs. The main rationale behind the volunteering computing concept is that computers often use around 10% or 15% of their computational resources so leaving a huge amount of resources empty.|
Thus, putting together several resources (laptops, desktops, etc.) into a unique, virtual computing infrastructure may substantially contribute to optimize some research projects day-to-day activities.
A current example of this trend is LHC@Home, located at CERN where your empty resources may be used for scientists to evaluate theory and real experimentation in areas related to the Universe creation.
|Another nice and old example is community networks (CNs). Indeed, CNs are defined as “telecommunication infrastructure deployed and operated by citizens to meet their own communication needs”. A very useful target for CNs is to provide connectivity in areas where is not accessible, whatever the reason may be. The seed for the concept may be found in the community memory initiative proposed in Berkeley in 1973. The key point is that the group of people sharing the resources is the one operating and owning the CN.|
Examples of CNs are AWMN (started back in 2002 in Athens and putting together 800 backbone nodes plus a few thousand Client nodes), Guifi.net (set in 2004 in Catalonia and being today the largest community network in the world with over 35,000 working nodes), TunapandaNET (in Nairobi) or Coventry and Warwickshire Network and Newham Online, both in UK.
Finally, another widely deployed example for sharing is PlanetLab, ruled by an heterogeneous consortium putting together industrial, academic and governmental institutions, to support the so-called PlanetLab overlay network. The network brings together around 1,000 widely distributed nodes all connected to Internet, setting a large infrastructure for network related research contributions testing and evaluation. Aligned to the “local” management procedure set in this sharing model, the PlanetLab network is managed by the PlanetLab consortium, responsible for updating the whole network, defining policies to rule the system and providing operational support.
Hence, it is evident that the benefits sharing may bring in different technological domains may be at least pretty relevant. But, certainly we can go far beyond that. Indeed, new scenarios started to pop up recently leveraging the sharing concept in different sectors and undoubtedly attracting great interest from the whole society. Widely used examples are AirBnB in the housing domain or several car sharing platforms in the automotive sector. Hence, the next step is obvious, may we extend this sharing approach to “new” user devices? Or in other words, why do not we benefit from the large disposition of ever more capable IoT devices and edge computing to facilitate the deployment of innovative services based on collaborative models? Certainly, devices at the edge are unstoppably growing in capabilities as never seen before, so if a normal utilization of a laptop or desktop leaves plenty of empty resources, why not extending that approach to smart phones or cars? In fact, this approach would be also aligned to the circular economy trend where resources utilization should be optimized. And even more, the yet to come deployment of 5G will unquestionably improve communication quality and capacity, so making connectivity to become a commodity. Thus, it looks like the perfect dream.
Therefore, it seems natural to start thinking on novel collaborative scenarios, definitely supported by the sharing concept, oriented to make the most out of novel smart devices, recently playing a key role in human daily activities, such as smart phones, smart cars or even smart city elements, with no need to forget traditional computing systems such as laptops or desktops. The main idea would push for going far beyond sharing a house or a car, towards sharing explicit system resources, such as computing, storage or even battery.
|Let’s for example assume a scenario of multiple (smart) cars parked in a garage and let’s assume these cars, when placed on the parking slot are doing nothing, so with plenty of empty resources. We may consider the whole set of cars in the mall as single computing units that may join all together to become a large computing facility offered by a “system manager”.|
Indeed, the owner of the garage may also act as a computing provider, setting an agreement with clients depending on what the client will is: i) only park the car; ii) park the car and share resources; iii) use computing resources.
This model may be applied to a hospital and in general to any place where cars may be placed for long. Moreover, this model may be also extended to include smart phones. For example, consider a big event concentrating thousands of people (concert, soccer, etc.). From a resources perspective the event is putting together an impressive set of resources, that when put together may easily become a super-computer facility, so why not using it?
That said, the theory is great, and the envisioned scenario looks promising. However, that may be not enough. If we move back to the house and car models, we easily see both scenarios meet a mandatory rule in the business sector, i.e., define the need and identify who is willing to pay for it. This rule is not so easy to be met in the envisioned collaborative scenarios proposed above. This assessment may be rooted on both, it is not clear what the benefits should be for a final user to share resources and is not so clear either who will be the provider for the whole integrated system. In short, if someone wants this interesting concept to move forward, a business model must be created, assuming the fact that the need and the resources are out there, but there is no one yet identified to properly manage (i.e., make money) the whole scenario.
 By the Internet Society at https://www.internetsociety.org/wp-content/uploads/2017/08/CommunityNetworkingAfrica_report_May2017_1.pdf