The Internet of Things (IoT) has quickly become the next “be all to end all” in information technology. Touted as how cloud computing will connect everyday things together, it is also feared as the real- life instantiation of The Terminator’s Skynet, where sentient robots team with an omnipresent and all-knowing entity that uses technology to control, and ultimately destroy, all of humanity.
Not there yetLucky for the humans among us, the technical capabilities of both cloud computing and IoT are way behind these Orwellian fears. Although the technology is promising, research and technical hurdles still abound. Challenges include:
- Datasets that span multiple continents and are independently managed by hundreds of suppliers and distributors;
- Volume and velocity of IoT dataflows exceed the capacity ad capability of any single centralized datacenter;
- Current inability to conduct “Big IoT” data processing across multiple distributed datacenters due to technical issues related to basic service stack for datacenter computing infrastructure, massive data processing models, trusted data management services, data-intensive workflow computing; and
- Benchmark limitations associated with heterogeneous datacenter application kernels.