Critical Thinking is a weekly column on innovation in data center infrastructure design and management. More about the column and the author here.
As we have discussed in this column before, AI could be a game-changer for data center management.
The hope is that AI-driven management software (likely cloud-based) will monitor and control IT and facilities infrastructure, as well as applications, seamlessly and holistically – potentially across multiple sites.
However, it’s still not exactly clear where this new wave of AI-enabled data center management tools is going to come from. There are several potential candidates, some of whom are already touting AI capabilities:
- Equipment suppliers. Infrastructure specialists such as Eaton, Siemens, Schneider Electric, and Vertiv are obvious candidates. Most of them have mature DCIM (Data Center Infrastructure Management) software as well as emerging cloud-based data center management-as-a-service, or DMaaS, platforms that will eventually be AI-enabled. The advantage these players have is that they have huge datasets to train AI and can also embed the required monitoring and control systems directly into their equipment.
- Large colocation operators like Equinix and Digital Realty have also developed in-house DCIM tools; they may look to extend these tools to include some form of AI in the future. Cloud services providers, such as Alphabet’s Google, are also developing their own tools. In fact, Google probably has the biggest field test of AI in data center management to date.
- There is also a whole raft of data center services companies providing everything from testing and commissioning to facilities management that are also investing in software platforms.
- And lastly, there are specialist software companies targeting this space that include suppliers of DCIM and Computerized Maintenance and Management Software, or CMMS, as well as a small but growing group of dedicated AI startups.
Last week, two companies representing the last two groups -- facilities management specialist CBRE and AI software startup LitBit – announced that they have joined forces to create what could be the biggest-scale application of AI for data center management yet – much bigger and potentially much more consequential than Google’s.
As Data Center Knowledge reported, CBRE currently manages more than 800 facilities on behalf of its customers. It has agreed to deploy LitBit’s AI-driven data center maintenance tools, which the San Jose, California-based startup describes as “Personas,” similar to Apple’s Siri or Amazon’s Alexa, to help automate some of its data center management processes.
Known as REMI, or Risk Exposure Mitigation Intelligence, the CBRE persona will be trained by and eventually work alongside the FM company’s staff. The hope is that REMI will eventually be able to detect mechanical and electrical infrastructure and environmental conditions on the data center floor and enable predictive maintenance.
Data Center Knowledge spoke with LitBit co-founder and VP JP Balajadia about the partnership with CBRE and the startup’s ambitions to improve data center management:
DCK: You already have an established data center persona called DACE, but the CBRE persona is called REMI. How are they different?
JP Balajadia: Really, what LitBit is trying to do is build a platform to enable people to become data scientists and creators of their own AI. The creation of DACE was more based on our internal thinking, as we have also been data center owners and operators. [Balajadia founded the data center provider 365 Main, and LitBit’s second co-founder, Scott Noteboom, ran data centers for Apple and Yahoo in the past.] We said, “What would we want to build into an AI persona to be a data center expert?” So, we came up with this fictitious persona that we wanted to be an expert in air conditioning, generators, pumps, etc. DACE was more of a framework for how our platform could be used. We also created frameworks for commercial real estate, and manufacturing, and petroleum, and oil exploration.
DCK: It’s quite unusual to hear from a cutting-edge AI software company that also has an understanding of the physical facilities side of the data center. Hopefully that knowledge is going to help overcome the fact that facilities teams can be quite conservative when it comes to new software and technology.
JPB: I always hate to hear people say, “This is the way we do things because this is the way we have always done them, and I don’t want to lose my job because I tried something experimental.” Some of the value that LitBit brings is that we allow them to scale as slowly or as quickly as they want. They can prove that value in a very non-invasive way. For example, introducing a DCIM system takes a lot of time and resources; we can do this via a mobile app and have your team training in a matter of days.
DCK: How does the commercial relationship work? Who owns the IP ultimately of REMI?
JPB: It’s a SaaS-product. As they add people and points onto the platform, the more pieces of gear, the more people that use it, and the more things they are doing, it creates AI work in our cloud. The measurement is work units – machine learning work units. That is the mechanism for the billing. Everything they put into the CBRE persona is theirs. We don’t own the data.
DCK: What could LitBit’s technology do for CBRE and its clients over time?
JPB: I think the way they see it right now is, “Hey we are a people-intensive company. People look at us and others in this industry and say we are manpower.” I think CBRE is thinking that over time it is going to transform into a technology company. It is going to leverage its human capital to build an AI that stays with the company forever. Eventually, the smartest person in the company is everywhere, because it is in the back pocket of everyone in every team. And it will never retire, it will never be hired away, and all the best people we have trained it collectively.