Issue link: http://cp.revolio.com/i/709701
THE DATA CENTER JOURNAL | 9 www.datacenterjournal.com Data centers are the core of today's business, and fiber-optic connectivity is the fabric, carrying vital data to drive critical business processes and providing connectivity to link servers, switches and storage systems. The physical layer F iber connectivity has become the most critical component in today's data center. Managing the physical- layer connectivity is mission critical. It's vital to ensure the continued five nines uptime of the data center. Data center managers have various tools at their disposal for monitoring faults in electrical, cooling and rack systems, but ideally, they want to manage their infra- structure by assessing the impact of adds, moves and changes on existing systems. If a manager plans to deploy 100 new serv- ers, for example, is there sufficient space, power and connectivity to accommodate them? A full-suite data center infrastruc- ture management (DCIM) system can provide such answers. Monitoring and managing a data center are not the same things, and DCIM tools make all the difference. MoniToring vs. ManageMenT In the vast connectivity array of the modern data center—with oen hundreds of thousands of high-speed fiber-optic connections—real-time monitoring of the physical-layer connectivity ensures that the data center manager is working with the right information at the right time. Monitoring is the collection of data, and managing is using that data to make better decisions about the overall data center. Monitoring is the entry point in improving data center management; it usually involves using a disparate set of tools such as intelligent power strips, intelligent meters, power-delivery units (PDUs) and busway systems to view the performance and operation of multiple pieces of equipment from different manu- facturers. For example, you might be using one vendor's data-aggregation soware for monitoring a busway system and another vendor's data-aggregation soware for monitoring power usage by your cooling infrastructure. ese disaggregated tools might tell you that the busway is working fine, or that the power strip has enough capac- ity. You might not know how the systems are working together, however. With a full DCIM suite, you can consolidate those disparate data points so you know that your cooling system is doing x, your busway is providing y to it, and your IT assets are consuming z in terms of power. You can, for example, correlate informa- tion to see what your IT load is doing to your cooling system, or vice versa. You can answer questions such as "Do I have sufficient cooling capacity to support ad- ditional servers in the data center?" Or "If my busway fails, what assets are affected? Will the IT assets power down, or will they keep running while my cooling system powers down?" Changes at the physical layer can have a profound impact, and the conse- quences of inadvertently removing the wrong connection can be catastrophic. With the growing complexity of con- nectivity (e.g., high-density multipath optical), intelligence at the physical layer is considered a must have. It's inconceivable nowadays to consider managing a modern data center without automated infrastructure manage- ment (AIM) systems, which provide intel- ligence at the physical layer. hoW dciM Works DCIM is a data aggregator. It uses open APIs and standard protocols like SNMP to collect information from many different systems and present them under a single interface. DCIM systems collect information on power, cooling, space and connectivity from the systems in the data center, and they work at the device, rack area and overall facility level. If you want asset-level (device-level) information, and it is exposed through SNMP or a data aggregator, the DCIM system can collect and display that information. Smart power strips are typically SNMP enabled, and if so, the DCIM can pull information from them individually or use data from your existing data-aggregator application. If you have bigger boxes such as an APS, a UPS or generators, that information is provided through a Modbus port, and the DCIM can query the systems to give you a clearer picture of what's happening with your power infrastructure. If you want to manage space and po- tential device adds, moves or changes, you can correlate data from different sources to get the complete picture. For example, if you have an intelligent power strip and you also have sensors that tell you which devices are in a rack, the DCIM can cor- relate the data and tell you what devices would be affected if the power strip goes down. Or you might discover that the rack has more devices than are actually run- ning, so you might pull some out of the rack to create more capacity. Most importantly, DCIM can use critical physical-layer connectivity infor- mation from AIM systems in real time to enable business-critical planning. is capability ensures that placement and deployment of assets and resources in the data center are underpinned by data that is both accurate and up to date. hoW dciM helps DCIM systems help you plan more effectively. If you want to know how much your data center can grow, DCIM can forecast and show you how much space, power, cooling and connectivity you have available. It can also show you which sys- tems are underused. For example, if you have 5 kW of power available to a rack and the rack only has three servers in it (two