If you think you've got a handle on how data centers work, think again. According to Gartner, the Internet of Things (IoT) is going to completely change the data center. The growth of the IoT isn't incremental; it's explosive. IDC predicts the percentage of data center capacity used by IoT devices will increase 750% by 2019.
That explosive growth means there are plenty of opportunities for value-added resellers (VARs) to work with a partner in order to identify ways the IoT can gain incremental revenue for them. Customers don't understand the impact of the IoT yet; VARs need to be prepared to walk them through changes to servers, storage, network, security, and other data center facilities.
Many of the changes that data centers need in order to support IoT have to do with providing additional capacity, but there are more fundamental changes that mean changes in corporate thinking about data center strategy. Customers are still adapting to cloud computing; now they need to start thinking about edge or fog computing.
And while the IoT will place additional demands on the data center, don't overlook the possibility of using IoT devices and capabilities in order to help manage data centers as well.
Changes in the Data Center in Order to Support the Internet of Things
Distributed, Not Centralized
There will be nearly 30 billion IoT devices by the end of the decade. One of the biggest changes needed to manage that many devices is a change in strategy. It isn't going to be possible to centralize that much data and computing. Expect companies to build smaller distributed data centers, where data can be aggregated and preliminary analyses conducted, with subsets of data passed on to other data centers for deeper analysis.
Data center designs will need to ensure their network has sufficient bandwidth. The number of connected devices multiplies a thousand times, and the traffic can overwhelm the wide area network. The nature of the traffic changes with IoT, too. There are frequent bursts of small amounts of data coming from all over—the IoT is "chatty." Much of the machine-to-machine (M2M) communication uses near-field communications protocols, and avoiding interference will be key.
Big Data Storage and Processing
All the accumulated data will need to be stored and processed. Many companies won't have the capacity to store the petabytes of data that IoT devices can generate and will have to rely on cloud storage. In addition, full backups of enormous IoT-generated data sets may not be possible or even desirable; companies will need to identify the data that have value and automate selective backup routines. Compressing and eliminating duplicate data will become a key part of data management.
Big data analytics methods are needed in order to process this data. Companies face two kinds of technical issues here. First, there's a lack of technical expertise in the required technologies, such as Hadoop. Second, some applications—especially those that extract, transform, and load big data sets—may be affected by latency on a network that's overloaded with M2M traffic.
Because of the volume of data, bringing it all into the data center—even a cloud data center—for processing may not be feasible. Processing will be pushed closer to the devices, at the edge of the network. Edge computing can greatly reduce the volume of data to be transferred, reduce the transmission time, and eliminate latency in receiving processing results. Cisco believes that 40% of IoT data will be analyzed at the edge (or in the fog) by 2018.
All of these new data need appropriate security, even while they need to be shared with partners or analyzed in combination with data from other sources. Particularly if the data isn’t centralized, ensuring encryption at the network edge is a critical part of security. In addition to data security, IoT devices create new network entry points and can make non-IoT data vulnerable. Because IoT interfaces offer control over machinery, there are physical vulnerabilities introduced by these devices as well, so data center security needs to address that level of concern also.
Companies need a way to ensure that only authorized devices are connected to the network. This means securely identifying and authenticating connected devices. Network devices need to use certificates in order to enable trusted connectivity. Companies can use software-defined networking in order to dynamically segment networks and control connections.
Data Center Management
With the increased number of devices and increased volume of data and network traffic come increased data center management issues. Using Building Management Systems (BMS) and Data Center Infrastructure Management (DCIM) software will help the data center operational staff understand the activity within its data center. It will become more necessary to use statistical throughput models when planning for future capacity needs.
Changes in the Data Center Supported by the IoT
Fortunately, the IoT contributes solutions as well as challenges to the data center. HVAC (heating, ventilating, and air conditioning) systems, such as smart thermostats, can help manage the data center cooling systems. Additional sensors can be placed in order to gather metrics such as power utilization and feed their data into facilities management software. VARs can look to incorporate IoT devices that benefit the data center besides making sure the data center design benefits IoT devices being used for other purposes.