The rise of the Internet of Things (IoT) is leading to an entirely new form of data infrastructure on the enterprise edge. This new layer will exist beyond the cloud in legions of micro centers designed to enable high-speed data services to an ever-expanding universe of connected devices. As such, this new edge will require a wealth of capabilities that do not, and cannot, exist in either the on-premises data center or the cloud facility, even as the basic technologies that drive the edge infiltrate IT infrastructure across the board.
Here, then, are 10 ways in which the edge will affect data environments end-to-end:https://o1.qnsr.com/log/p.gif?;n=203;c=204663295;s=11915;x=7936;f=201904081034270;u=j;z=TIMESTAMP;a=20410779;e=i
Connected devices require continuous support and immediate response to data requests or, literally, as in the case of autonomous vehicles or medical devices, lives could be put at risk. This is one of the reasons why the edge exists in the first place: to push processing centers close to users. But it also requires the latest in processors, storage and networking, all of which should migrate into centralized data infrastructure simply as a means to improve cost/performance ratios.
Less Processing on the Cloud, More Storage
Still, with this new layer of processing on the edge, it seems likely that it will soon take on the critical workloads that drive business models. This isn’t to say the cloud won’t be used for analytics and back-office processing, but in terms of its role in overall enterprise workloads, it is possible that the cloud will largely revert back to its original value proposition: as a means to support highly available storage at low cost.
Global Data Footprints
Connectivity is expected to permeate virtually everything we touch and see, perhaps even our own bodies as well. This will push the enterprise data environment to a global scale faster than any technological advancement that has come before. With ecommerce already making it possible to sell any good or service to anyone, the IoT will make it possible to maintain steady streams of data virtually anywhere, even as users start venturing into space.
Since much of the IoT edge will be unmanned, the need for thinking machines that can devise their own solutions to tricky problems will become paramount. As well, technologies like neural networking, natural language processing and facial/image recognition are expected to make interaction with devices easier and more intuitive, which in turn will drive increasing levels of intelligence both on the edge and throughout the enterprise data chain.
The edge is where the worlds of wired and wireless collide, so it will require the latest in 4G/5G capability as well as traditional HTML and both long haul and local networking support. The rise of mobile technologies has already brought a good deal of integration between wired and wireless, but the IoT kicks it to a new level in which the two will function as a single network.
The key operational paradigm of the IoT is that everything communicates up and down the data ecosystem: devices, systems, platforms, protocols, hardware, software… This means a high degree of interoperability is required, most likely driven to a large extent by open environments. As interoperability permeates one layer of IT infrastructure, we can expect it to impact local and cloud architectures as well.
The micro data center will have to be compact by necessity. This will require extremely modular, hyperconverged, hardware that can be easily assembled and replaced when it becomes necessary. As with interoperability and advances in processing power and storage capability, convergence will also start to take over traditional data infrastructure, most likely to the point where all physical layer technology is commoditized in support of highly flexible abstract data architectures.
The hierarchical structure of both block and file storage is wholly unsuited to the massive scale and highly dynamic nature of the IoT. This puts object storage in a prime position to take on the bulk of edge storage infrastructure, and in turn much of the centralized storage the supports less-time-sensitive workloads. Object storage is also more amenable to HTTP and RESTful APIs, which happen to support the bulk of IoT applications.
In traditional environments, the enterprise is under pressure to save all data, just in case. But this is impossible in the IoT. Instead, organizations will have to rely on advanced analytics and finely tuned governance policies to determine what goes and what stays. Fortunately, much of the data generated by devices is very ephemeral in nature, which means its value can drop to zero in a matter of seconds.
Thousands, perhaps millions, of data streams all clamoring for attention at every moment: This looks like a job for massively parallel processing. Expect the edge to incorporate the latest in GPUs and multi-threaded interconnect architectures in order to handle these massive volumes. And again, what looks good on the edge will start to look good for the data center as well.
Arthur Cole writes about infrastructure for IT Business Edge. Cole has been covering the high-tech media and computing industries for more than 20 years, having served as editor of TV Technology, Video Technology News, Internet News and Multimedia Weekly. His contributions have appeared in Communications Today and Enterprise Networking Planet and as web content for numerous high-tech clients like TwinStrata and Carpathia. Follow Art on Twitter @acole602.