How to get control of large numbers of data center alarms.
By Don Klein, Modius Inc.bo
Data centers continue to grow tremendously in size and complexity. Individual data centers can cover many acres, frequently spreading over multiple sites. To illustrate, the Lakeside Technology Center in Chicago is 1.1 million square feet and has 50 backup generators. Data centers of this size hold thousands of racks, tens of thousands of servers and dozens of precision air conditioners. No team can manage this much equipment by relying on physical senses, listening for sounds and looking for lights. Requiring physical proximity to managed devices is a non-starter.
Regardless of the total square footage of a data center, the only scalable approach to information technology (IT) and critical facility management involves having IT devices and the systems that support them send notifications to a centralized management system that can distill meaning from the noise. Normalizing, correlating and prioritizing these disparate messages to generate a coherent representation of the state of the data center is the first step in maintaining control over critical IT infrastructure and services. This single, consolidated view of the status of all the alarms in a data center is called unified alarm monitoring.
Unified alarm monitoring provides exceptional benefits to operators of data centers and critical facilities. Such systems possess the following characteristics.
- Monitoring of power, cooling and environmental systems’ responsiveness to changing conditions
- Management information base (MIB) polling to gather data before alarms are generated
- Aggregation and correlation of data into “meta-alarms”
- Precision in identifying the location of a problem
- Prioritization of multiple issues
- Escalation of issues to responsible parties
- Support for heterogeneous IT environments and multiple protocols
- Integration with third-party systems
- Simplicity and ease of use
With the right monitoring system, operators at a single command point can see and respond to all excursions in the condition of IT equipment.
To be sure, the processing power necessary to aggregate large quantities of alarms at one monitoring location can be expensive. One approach that managers of data centers, control rooms and other critical facilities can take to reduce costs is trying an entry-level system. The limited systems help enterprise senior management achieve the initial business benefits of unified alarm monitoring without breaking the bank.
We have found three types of entry-level systems. Each is different and each appeals to a different cross-section of users having distinct interests and needs.Basic alarm collection systems reduce cost by aggregating the alarms and displaying them on one screen.
Remote alarm alerts inform data center managers wirelessly about alarm conditions. These wireless alert systems appeal to data center managers who are frequently away from the data center, or are out of the critical facility but nearby for extended periods. These techniques also allow for remote alarm monitoring overnight or during holiday periods.
Software as a Service (SaaS) system monitoring products can be used to monitor widespread networks. They excel at controlling extensive systems covering locations at numerous sites. They are purchased by operations managers who value dashboard presentations.
If an entry-level system is chosen, it is important to pick one that can be upgraded later to provide a richer feature set as needs evolve.
The first step for data center managers who want unified alarm management is determining goals and criteria for alarm-monitoring solutions. This deliberation should involve representatives from corporate, IT and facilities. Once the team has clarified its needs, the project managers should compare the major alternatives. Finally, the data center team should purchase or lease a system with the capabilities they believe will do the work they want at a price they can afford.
Unified alarm monitoring is essential to maximizing the availability of your data center’s critical services.
Don Klein is vice president of marketing with Modius Inc. (www.modius.com). Modius provides performance-management software systems that are used to monitor infrastructure within data centers, and to analyze data center performance. The company’s product line includes the capabilities that Klein described in this article.
Past CIM Issues