More feature stories by year:
Return to: 2018 Feature Stories
Feb. 20, 2018: Data Center Post
Although many enterprises have adopted data center infrastructure management (DCIM), still more have not made the jump. DCIM ushered in improved operational efficiency backed by service level agreements, transparency, and reporting. But there remain areas of DCIM that could be enhanced to gain broader adoption.
Few doubt the many inherent benefits of DCIM. Applied correctly, DCIM solutions save time and money, reduce manpower, and enhance human productivity. A blessing to data center managers, DCIM tools obviate the need for employees to be on-site to identify assets in place, the space assets occupy, or the power and cooling resources they consume at any given time. Managers can access accurate, real-time data from laptop or desktop. DCIM gives them a window into key asset details, including physical power and network connections. A boon to accurate capacity planning, DCIM tools help identify critical path capacity points to cut failure risks down to acceptable levels. Reservations, moves, adds, and changes go smoothly. Even employee productivity and morale go up when DCIM goes in, thanks to better processes and workflows.
Many data center managers give DCIM a passing but not excellent grade. The reason for this is simple: deploying DCIM remains a challenge for many enterprises. Some have even sidestepped commercial DCIM products and built customized versions to suit their specific needs.
Besides being a software product, DCIM involves a process that draws on multiple groups within an organization–typically facilities, data center and IT. Implementing DCIM solutions means working face-to-face with people in distinctly separate areas of an organization. In doing so, managers must define the processes to be managed via the DCIM solution and ensure that data currently available is indeed accurate. This means doing the grunt work of setting up SNMP community strings and Modbus registers, as well as IPMI usernames and passwords to poll data from the “gear” in the data center. Equally important is working the DCIM solution to gather data from legacy hardware using existing and often undocumented protocols. Getting up to speed on a DCIM solution means doing all the “legwork and linework” to determine a data center’s unique needs, to evaluate possible solutions, and to schedule a “proof of concept” to verify that the solution chosen meets the needs in a data center’s environment.
Datacenter Management as a Service (DMaaS) emerged in the second half of 2016. What set it apart from DCIM tools was the collection and analysis of data from a variety of clients with diverse data centers in widespread locales. The “big idea” for CIOs was the ability to apply big-data statistical analysis. This was spearheaded by the promise of machine learning and other forms of AI to extend DCIM’s value reach, especially with predictive features. The statistical analysis benefits of DMaaS are being embraced by data center and IT managers as investments in on-premises DCIM platforms can now be moved to cloud services.
DMaaS finally makes it possible to accurately predict and prevent data center infrastructure incidents and failures. Inefficiencies and capacity shortfalls are set in sharp relief under DMaaS tools. This cloud-based, vendor-neutral software-as-a-service architecture moves data centers, edge, and hybrid cloud into the future. In contrast to on-prem and SaaS-delivered DCIM, DMaaS aggregates and analyzes opted-in, anonymous data that can be enhanced with machine learning. The value added is that DMaaS unites cloud-based monitoring with maintenance and repair services, a boon to suppliers.
Not going unnoticed is that fact that DMaaS is easy to adopt, allowing baselining and benchmarking against similar data centers. Its automation and monitoring benefits appeal to IT and DevOps purists, who prefer easy and instant access to vast silos of collected data. Gleaned from the “global cloud,” these data analytics form the fulcrum leveraging the key advantages of machine learning and increased knowledge within the cloud. With DMaaS, data can be accessed anywhere, which means managers needn’t be on site or rely on VPN to assess a risk or add the right people to address a problem.
It may still be a few years away, but cloud-based AI-driven management software will soon seamlessly monitor and control IT and facilities infrastructure, including applications across multiple sites. AI is the genie in the bottle that will oversee control functions, and monitor/adjust cooling, power, computing, workloads, storage, and networking to optimize efficiency, productivity, and availability. Cloud-based analytics will use sensor data from multiple sites to guide and implement preventive maintenance programs. Robots will order, test and install spare parts as needed to reduce failures and virtually eliminate unnecessary maintenance and testing.
As DCIM continues to grow in adoption, data center managers will be looking to the next step up in operational efficiency. DMaaS expands DCIM with bigdata statistical analysis and machine learning. It allows investments in on-premises DCIM platforms to finally be moved to cloud services.
Return to: 2018 Feature Stories