Everyone is talking about mobile applications and “e” or “m” enabling services. Hackathons and developer competitions are the order of the day with many looking to create the next Instagram. While these types of conversations are taking a large percentage of local tech’s share of voice, an interesting initiative is slowing taking root that is looking at a different aspect of tech that when successful will open up new opportunities for both business and learning, with its different approach to what I call “capacity as a service”.
A team of local developers with some financial backing from Google and working out of the Ihub are creating what in my books is Kenya’s first locally assembled supercomputer. The project has been a while coming and conversations around it have been interesting with a CEO of an established tech services firm quipping “ Why do we even need a super computer? ” This is expected as the majority of client facing deployments of technology services are not capacity intensive unless you are running a Facebook of sorts.
A super computer refers to a high performance computer setup, built to deliver immense processing power. While the local deployment may not rank anywhere close to the builds deployed the world over, with costs in the millions of dollars and used by military, pharmaceutical, meteorological and manufacturing outfits, it is still an important first step in what I see as a more tangible form of the outsourcing business that has proved elusive.
On the education front, the team is documenting the project build to allow students and other enthusiasts to learn how to build hardware infrastructure; breaking down existing hardware setups and reconfiguring them to meet specific needs and thinking through holistic platform provisioning. Academia would do well to pay attention to this or even incorporate it into a module or two; the hands on nature of this project offers rich learning.
On the commercial side, the applications are many with the predominant model as capacity sold per hour. Government can blast through tonnes of data generated by is operations daily to derive insights to guide decision making, architects can output their plans with ease, scientists can see accelerated discovery from their research work, local councils can power smart cities, process intensive software applications or mobile applications targeted for the masses can run more efficiently and with the burgeoning content industry, animators can also render their work within the shortest possible time.
High performance computing can make money by doing the heavy lifting and thereby improving efficiencies across all sectors of the economy and this initiative will be one to keep an eye on; the spin-offs may just prove to be interesting.