Data center technology vendors usually offer remote monitoring or management tools for their products. They've also gotten better at providing APIs for easier access by other systems. But there’s still a missing piece: a platform that can pull everything together into a single control plane.
"And that has been a problem in our space for as long as our space has been around," Brent Bensten, CTO of products at the data center provider QTS Realty, said. APIs help, but they are just part of the answer.
QTS had to build its own integrations. "A lot of the work we're doing is interfacing with other tools,” he told DCK. “It's 100 percent siloed."
It’s a problem because technologies used in a data center are interdependent. Plus, any meaningful addition of AI capabilities into the environment would require a comprehensive view of it.
"Until you can amalgamate multiple different sources, in [my] opinion, it's a solution, but it's not a holistic solution that's as powerful as it can be," Bensten said.
Take, for example, DCIM software platforms. (DCIM stands for Data Center Infrastructure Management.)
"None of the tools in the DCIM tier support the concept of service providers," he said. "Therefore, multitenancy causes them to bend – if not break – before you even start."
Then there are all the other systems that are not part of core DCIM, such as access systems, water usage systems, and video systems. According to Bensten, none of these tools are designed to interoperate with each other, providing no value outside of their single design purpose.
Pulling data and logs from systems is just the first step. When data centers are looking beyond simply monitoring and observing disparate systems, to managing them, the challenge gets even more complex.
"They all have individual interfaces," Constellation Research analyst Holger Mueller told DCK. "The story is different, of course, if you run 100 percent on one stack – for example, all on VMware, or all on HPE – but almost no real company does that."
According to Daniel Newton, CEO at CDS, which provides technology services to data centers, numerous technologies provide some of the integration and overview that data center operators are looking for.
They can start with DCIM software and network performance monitoring and diagnostics (NPMD) and evolve to IT operations management (ITOM) and application performance monitoring (APM). "In addition, security information and event management (SIEM) is now table stakes," he said.
The current pandemic environment has put an additional spotlight specifically on data center capacity management, he added. "Organizations need to be focused on deploying management and monitoring solutions that can cover as much of the data center as possible.”
That minimizes the need to deploy and integrate multiple different solutions and aggregate multiple different datasets, Newton said. "As far as possible, the solutions deployed should be out-of-the-box and plug-and-play with API capability in order to avoid the need to acquire new team skills and capability."
Sunbird's DCIM platform, for example, supports standard protocols and APIs, said company president Herman Chan.
DCIM software is typically used to monitor physical data center infrastructure, such as heating and cooling, capacity and usage, and rack or cabinet space. It's important to have this in one platform, said Chan. Deploying too many virtual machines on servers, for example, can put a strain on power systems, trip circuit breakers, and cause downtime.
"Users need vendor-agnostic solutions that are open and interoperable," he said.
With the pandemic, DCIM systems have become the most in-demand tool, he said. In response, Sunbird offered free remote monitoring for the healthcare industry and free weekly training to customers.
DCIMs and SIEMs have limitations. One example is they interface only with on-premises systems and equipment. But the logjam may be starting to break. According to Mueller, some of the big players are beginning to cast an eye at this market.
One vendor trying to address the problem is Oracle, with its Oracle Cloud Observability and Management platform. The system, released just a month ago, gives visibility into multi-cloud environments, hybrid environments, and on-premises infrastructure.
It's one of the first products that addresses both the vendor's own technology stack and third-party platforms, both on-premises and in the cloud, said Mueller.
"Google is also working on it," he said. "It is just starting.”
Last spring, for example, Google announced availability of Anthos, a multi-cloud application platform built around Kubernetes that can be applied in the cloud or on premises. But it’s unknown whether Google has any plans to add visibility beyond the servers it runs on and include things like power, temperature, and humidity monitoring.