benedetti68 - Fotolia
DCIM is moving from an on-premises design to a cloud-based approach in data center management as a service, but DMaaS products still face an uncertain future as they go up against on-premises offerings.
Research firm IDC expects that companies will spend $624 million on data center infrastructure management software in 2019, a 10% increase from the previous year. Halfway through the year, those numbers are steady, but still below expectations from a few years ago.
A reason behind these low numbers is that there are some adoption hurdles with DCIM products. The tools can be difficult to deploy and use. DCIM software must be installed on every device in a data center, and it requires a great deal of customization to parse the collected information. Also, the first wave of software was built for a single data center and did not easily scale to support multiple sites.
This setup changed with the emergence of cloud technology.
"Cloud forced companies to examine how they deploy and maintain data center equipment," said Jennifer Cooke, research director at IDC. As a result, companies are moving workloads out from their data center to third-party sites.
A cloud option emerges
So, cloud-based DCIM, deemed data center management as a service, emerged. DMaaS offerings started arriving in late 2016, said Rhonda Ascierto, vice president of research at Uptime Institute.
Vendors on the market with DMaaS offerings include Nlyte Software, Schneider Electric, Data4 and NTT Communications.
DMaaS has an updated architecture and can be easily deployed across any cloud-connecting infrastructure. These services support a variety of devices and user experiences, such as smartphones and tablets, so admins only need an internet connection to gain instant access to the software and remotely monitor their infrastructure.
Power in data center management as a service
With cloud-based infrastructure management, vendors collect system usage information from many customers instead of one, isolated company. Admins then apply the vendor's AI and machine learning engines and get insights into areas such as power needs, energy costs, cooling efficiency, asset management and carbon emissions.
"Predictive maintenance is an area of emphasis with the cloud systems," Cooke said.
Corporations now update their energy devices and server workloads on set schedules, which can be an inefficient use of hardware. With analytics, admins see how the infrastructure is working currently and use the information to hone operation schedules. This allows IT teams to save energy costs, run a more efficient data center and update the right hardware.
The cost question
But DMaaS has limitations. With software-only systems debuting on the market three years ago, these offerings are in an early stage of deployment. In some cases, they do not offer as wide or deep a range of functionality as hardware-based systems.
Cloud is impractical in some cases. Sending alarm data to and from a centralized facility creates an unacceptable delay for time-sensitive applications, such as real-time alarming. Organizations might also not be able to use cloud-based infrastructure depending on their use cases or security concerns.
Cost is a convoluted issue. Initially, cloud vendors offered these services for little to no cost or as beta features. This way, vendors could gain customers and train any AI systems.
However, suppliers now must modify their pricing models and turn a profit.
"DMaaS pricing plans are all over the place," Ascierto said.
The main metric for billing is usage, though vendors vary just how they track resources. In some cases, vendors charge per device, per year, per month or per day.
Another option is to focus on usage minutes, which can be calculated through software time tracking or resource uptimes. Sometimes, vendors take a percentage of the savings that corporations attain on their energy bills.
Software capabilities are another component to pricing. Tiered pricing is quite common. Vendors offer broad, high-level reports for free and then charge for more detailed or customized ones.
Too much clutter?
The DCIM landscape is changing, and companies are hesitant to install multiple management systems, because it can add to software complexity. DMaaS requires that they add one more tool to their data center toolbox.
Another pricing and configuration option is having vendors roll DMaaS capabilities into their product suites. Corporations that are already using the software may opt for their DMaaS add-on and avoid adding management system interface clutter.
Analysts expect wide adoption of these bundles. "In the future, DCIM will not be the high-level interface that for the CIO relies on, which the vendors thought a few years ago," Ascierto said. "Instead, [the data center] will feed its information into that system."