5 Budget-Friendly Ways to Tighten Up the Data Center
In the fast-paced world of e-commerce, there are few things more important than the availability and performance of corporate applications and data. Many businesses do not realize that such a small investment to properly equip and maintain their data center can reduce so much risk in the future. Points like fire protection, cooling and physical security must be addressed.
Mar 30, 2011 5:00 AM PT
As online shopping becomes a larger and larger part of global consumers' spending habits, they are demanding the highest levels of availability, performance and security. Your application cannot be unavailable, slow or vulnerable. Period.
The single most important factor in keeping an application live and performing well is ensuring that your data center is running properly. Maintaining an up-to-date, secure and safely redundant data center seems like an expensive and tedious order, but there are actually many budget-friendly steps that can be taken to improve operations.
Two primary rules guide the most efficient data center operations. The first is this: Buying cheaper alternatives early on will always cost more in the long run. Data centers are inherently expensive, but cutting corners with critical systems will create far greater expenses in both repairs and the customer attrition that will inevitably occur.
The second rule is very simple and doesn't cost anything: Create checklists! Creating operational checklists and following them on a daily, monthly or annual basis will make a huge difference in equipment longevity.
There are five critical areas to any datacenter's operations: power, cooling, fire suppression, security and network. These are mutually dependent functions; all of them need to be running smoothly to avoid downtime.
- Power is obviously a critical component, and all redundancies should be inspected carefully. UPS and TVSS systems are critically important and should be checked daily unless connected to a centralized monitoring system that can detect and alert you to anomalies. Generators should be exercised on a biweekly basis to make sure that they are capable of holding the load of the data center.
Monthly checklist items include checking voltage, amperage and balance of three-phase power at each major point of the distribution system. In addition, all electrical circuits should be measured for temperature and inspected on a regular basis, and the 80 percent usage rule should always be adhered to.
- Cooling is essential, but many data centers are spending far too much money on it. One simple consideration that can make a huge difference in budget is proper airflow. Air mixing diminishes efficiency when managing cooling, so supply and return duct locations should be carefully engineered. This can be as easy as moving a perforated floor tile or popping out a ceiling tile to move a flexible duct. Proper airflow will without a doubt make your BTUs go much further.
Cooling is another area where monitoring plays an important role. Temperature, humidity and air flow should all be easy to allocate with modern CRAC units, but alerts should also be set up from these devices to operations personnel via mobile devices. Complications such as water leaks within the unit can cause a great deal of damage if not resolved immediately.
- Security is a more expansive process that must be carefully thought-out by any data center. To mitigate physical intrusion threats, high-end datacenters can install vibration detection in the walls and floors, as well as motion sensors in all ceiling cavities. Other more immediate measures include video surveillance of the facility and sufficient length of video storage in case an abnormality is not noticed immediately. Dual-factor access using proximity cards along with a fingerprint or hand scan is essential.
Human resource vetting is perhaps even more vital than these physical security measures; and background checks, including financial history, should be conducted on all facility personnel. Network security requires protection against DDoS attacks and other types of network intrusion.
The equipment is expensive, but it's budget-friendly when compared to losing business caused by a serious network outage.
- Fire suppression is a function that should be planned for both proactively and reactively. Obviously, combustibles should never be allowed on the server floor. With quick response in mind, VESDA systems are critically important to respond to threats as early as possible. Clean Agent fire extinguishers should also be placed within easy reach on the data center floor if not installed as a complete system. This chemical is affordable, non-conductive, non-damaging to servers, is not harmful to humans and evaporates quickly, and is thus an obvious choice for data centers.
In the fast-paced world of e-commerce, there are few things more important than the availability and performance of corporate applications and data. Many businesses do not realize that such a small investment to properly equip and maintain their data center can reduce so much risk in the future.
If all five areas of datacenter management are carefully designed and monitored, your datacenter will run like a top, and your customers will receive greater performance and availability than they could have imagined.