What do data center operators really know about the IOT and what it can do for them? More precisely, what can the Industrial IOT do for them. I’d like to draw this distinction, because the use of IOT has been well documented by many of the analysts (Gartner, 451Research and others.) Those reports and analyses center around the use of the IOT in relation to the IT systems that make up the data center: networking, storage, memory, compute and security…important topics, but ground that is well covered. The IOT will make those pieces of the puzzle faster, cheaper, nimbler, more efficient, more predictable and (subject of many debates) more manageable.
The industrial devices that provide the ‘data center’ environment needed to keep these boxes running can also benefit from the use of IOT - and have been for decades. Data center and Industrial infrastructure operators have been using a lot of what defines the IOT, before the IOT was cool. In 1979 the Modbus protocol was introduced to the industrial controls market as an open sourced communications platform to connect intelligent devices. In 1999 the Modbus TCP/IP specification was developed to move Modbus into the 21st century and allow it to be connected to the internet. The first mention of the Internet of Things was in 1999, a pretty interesting coincidence.
A problem that all data center infrastructure operators have (myself included) was that the controllers used are expensive and slow by modern standards. (A future blog post will cover the topic of the slow pace of innovation by the industrial controller suppliers.) The second, and most important problem is that the data created by these systems is locked away in the facilities engineering silo and only accessible by a small group of people. The real power of the IOT is the ability to take information and share it among other groups that can gain insights from it.
How much better could efficiencies be…how much more manageable would systems and facilities get, if all of that information was shared with everyone? I think that first we need to dispel any myths that we have about the IIOT and data.
The IIoT is not about the internet and it is not about things…it is about data…big data…lots of BIG DATA. We need to stop fearing big data and embrace it for what it can ultimately do for us. The easiest way to start using it is to understand what it is and what it can do:
- Data is not just about the numbers
- Data is not equal to truth
- Bigger data is not better data
- Data does not kill innovation
Data should act as a map to discovering insights about what you are measuring and how those insights might connect things that might not have occurred to you and your teams. If it was just about numbers people would hire more mathematicians than data scientists. Likewise, access among different groups will allow new insights to be discovered, new processes to be written and new efficiencies to be gained. “What if…?” Should become the question all teams ask each other without fear.
Data has biases. The most common biases we will find in our data will be signal bias and algorithmic bias. Signal bias may be due to improper or limited inputs: not having the right sensors in the right place or too many sensors in the wrong place. Algorithmic bias is due to data analysts skewing results in the wrong direction – i.e. the one time I searched for a replacement battery for a Sea-Doo lead to continuous ads for SCUBA equipment and dive vacations all over my browser. The goal should be to ignore biases in the data and to treat it in a more Bayesian way – using more probability and statistics and insights from people that operate the systems rather than making blind decisions based on data.
Bigger data is not better data. If you have a sensor providing you with information thousands of times a second that never changes (i.e. a thermostat reporting that the temperature is 75-degress…1,000 times per second), that is big data…but it is useless data. Having that information streamed from the devices to servers and to the cloud is not only useless, but a huge waste of resources. People need to provide inputs to unlock the potential of the data, it needs to be personal and it needs to be meaningful. Different people will find different data meaningful so the ability to share it among different groups is a must.
Data should be used to innovate and can be paired: data and discovery; data and improvement to come up with new and innovative solutions. How would you operate a data center if you knew that the compute load was going to be low all day and the server room could have its temperature increased by a few degrees and not hurt anything? How much smarter would your company be if data from one building was shared instantly with all of the others in the portfolio? How much happier would your customers be if they could access their rooms data on their mobile devices?
These are problems we are working on with our customers, providing them and their teams with access to real time data on desktop and mobile applications, the ability to sift through the data for meaningful insights, the ability to use machine learning to help systems operate more efficiently and automate tasks and the ability to share that knowledge base with their teams and the community we are building at Litbit. It is great to learn how people use this information once it is unlocked and given meaning.