IT Infrastructure and IoT: Data Centers, Cloud, Branch Offices and everything in between
By: Marcio Saito, CTO, Opengear
“We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten.” Bill Gates
The Internet-of-Things (IoT) is a set of technologies that allows intelligence, sensors and actuators embedded in physical devices to internetwork to collect and exchange data. As computer technology miniaturises, we can envision a world where objects around us are intelligent and connected, transforming how we interact with information and technology.
The implication for the IT infrastructure behind IoT is the possibility of a very large volume of data to be transmitted, sorted, stored and analysed. There is ongoing discussion on how centralised that infrastructure should be. Do we funnel and transport all data to the core or deploy components to aggregate it closer to the edge?
In this article, we explore the answer to that question, which is relevant to the strategy of implementers and users of IoT technology in the enterprise space.
The IoT Hype Cycle
IoT is a transformational wave that has been in movement for a few years and will cause ripple effects for decades.
As the quote topping this article suggests, experts and users alike have a hard time evaluating the impact of such large shifts in technology. We tend to overestimate the change that occurs in the short term (we get excited too early) and then, when the hype subsides, we underestimate the eventual effects of new technology (we get frustrated, dismiss the trend as a fad and fail to benefit from it).
Being able to see through the hype cycle and position correctly is what separates those who are going to be successful from those who are going to fail as consequence of change.
The early stages of IoT adoption in the Enterprise
While IoT technology has been evolving for several years, we are still in the early stages of adoption by the enterprise. Most use cases involve the leverage of mobile/wireless technologies, but still lack the very large number of edge devices anticipated in most IoT predictions.
Let’s take the example of a retail store. Until recently, there could have been a dozen Point-of-Sale (POS) cash registers, and a couple of desktop computers at the front of the shop used by managers to access the inventory system. Perhaps there were a few security cameras. Those devices connect to servers in a data closet running the commercial automation, inventory management, video storage applications. Data was batched and uploaded daily or weekly to some server over the corporate network.
As we walk into a more modern retail store today, we see cash registers replaced by tablets equipped with a credit card reader and running a POS app. Store managers use scanners and other handheld devices to collect and access inventory data. There are security web cameras. Those devices are often managed through a smartphone app and connected through a Wi-Fi access point directly with servers in the cloud. PC computers and closet servers are gone.
In this scenario, there is a migration of complexity from the edge of the IT infrastructure to the core. The link between the remote sites to the Internet becomes more critical, as each device now is dependent on its connection to a server in the cloud. Ensuring the Resilience of your network is critical to support the distributed business. For implementers of enterprise IoT, that means a focus on connectivity and need for Failover strategies to cope with link failures.
IoT Data Explosion
As the technologies mature and the enterprise use cases of IoT multiply in the next years, we will eventually see explosion of data volumes generated and consumed at the edge of the network.
Let’s continue with the retail store example. Hundreds of tags indicating the price of merchandise on the shelves will be replaced by smart tags that can be remotely controlled. Intelligence embedded in shelves and the products themselves will enable monitoring and management of inventory levels through a centrally deployed application, without human intervention. There will be sensors and beacons distributed throughout the facility to control and monitor everything, ranging from the customer movements and environmental conditions, to physical security, etc. Streams of video content might be displayed and recorded continuously.
In that scenario it may be necessary to introduce new IT components closer to the edge of the network again to aggregate, filter, sort and analyse data to keep it from unnecessarily flowing to the core of the network. Those components might include mesh wireless gateways, data aggregators, local data storage, and analysis engines. In large IoT deployments, it might be necessary to deploy small data center-like infrastructures closer to the edge of the network. With more IT components at the edge, Remote Provisioning and Out-of-Band management technologies will be needed to minimise the cost of maintaining the IoT infrastructure in operation.
With increased complexity, scale and dispersion of the IT infrastructure needed to support IoT applications, Automation also becomes an important component of the IT strategy.
Most analysis pieces on consequences of adoption of IoT in IT Infrastructure will focus on the end scenario where the technology is adopted and the use cases are fully developed.
We suggest that, in a large development like IoT, it is important to see through the hype cycle and note that this movement happens over a decade, not months. From the IoT adopter’s perspective, being aware of that can mean the difference between success and failure.
For most enterprise users, adoption of cloud and mobile technologies and a focus on Network Resilience is the first step. As IoT adoption matures, the need to build additional infrastructure closer to the network edge and focus on provisioning, remote management and IT automation will become more important.