There are interesting notes from Gartner’s Infrastructure, Operations and Data Center Summit 2014 held on May 21-22 in Sydney. Quoting directly from Fabrizio Biscotti, research director at Gartner, “IoT deployments will generate large quantities of data that need to be processed and analyzed in real time.” He goes on further to say “processing large quantities of IoT data in real time will increase as a proportion of workloads of data centers, leaving providers facing new security, capacity and analytics challenges.”
To put it simply: Internet-connected devices will transmit real-time data to central data repositories (centralized management systems or databases if you will) which will be analyzed by the big data infrastructure to understand how the business is functioning, and to allow the executives make informed decisions.
Internet of Things Example: Tesla
Wire reports that The National Highway Traffic Safety Administration has published two recall announcements, one from Tesla Motors and one from GM, both related to problems that could cause fires. Tesla is fixing the situation with an over-the-air update, without requiring the owners to bring their cars to the dealer. If it sounds like updating your smartphone or tablet, it is. Your smartphone is connected to your wifi at home and sitting idle, your car is in your garage/parking lot and is sitting idle. Why not fix it or add new features to it?
This data connection means that you can also receive data in real time. Consider Nest, where Google wants to take a hold in home automation and meantime collect data. Or consider Toyota, where data flows from all its authorized dealers/repair shops for all Toyota vehicles under repairs/maintenance. As Red Hat officials explained it in Red Hat Summit 2014 Istanbul, that data Toyota handles is 200 Gigabytes per week!
It is no-brainer that such a flow of data and the computing power required to handle this data bring challenges to the data center. Such challenges can be summarized as follows:
- WAN: Current wide area links are optimized for human to human or human to machine interactions such as emails, chats, audio/video conferences. IoT dramatically challenges this flow with the machine to machine interactions with the flows of sensor data. The WAN links should be reconsidered for the inbound flow of such a massive data.
- Servers: The analysis of such massive information will definitely require rethinking of the servers in the data center, perhaps leading to the acquisition of purpose-built servers.
- Storage: It is also a no-brainer that storage space will need to be expanded to accumulate the data.
- Security: Data is the blood of the companies, the flow of data means the veins. Not securing the veins means that the company is bleeding, to say it is leaking its corporate secrets. The data must be secured both geography-wise, in rural areas, in urban environments, and access-wise, who gets access to what information.
- Privacy: The data coming from the sensors is not only about the device itself but also about how the owner uses the device. Following the example above, Tesla receives information about the car itself and the driver. The challenge is how to protect the driver’s personal use information: driving information such as acceleration, braking, and equipment use such as navigation equipment, in-car hi-fi system etc.. How will all these will be secured?
One Data Center to Rule Them All?
It is not likely that today’s centralized data center will be able to receive, store, process, catalog and analyze this vast amount of information. Focusing on such a centralized data center is incompatible with IoT, neither in economic nor technical points of view.
Corporations who will be using IoT will be forced to build smaller data centers which collect and initially process the raw information, which then forward the relevant information to the central site for additional processing and analysis.
If the companies use IoT as we pictured in this article, the companies should also be thinking about the Data Center Infrastructure Management. Transforming the data center only will not be enough.
Image credit: datacenterknowledge.com