Data Center Operation Management

Key points in Data Center Management to ensure improved up-time:

• Ensure network & power cables are well dressed to ensure proper cold air flow from front of device and release of hot air from back side without obstruction. Gradual heating and prolonged higher temperature of electronic components inside device will reduce their life and overall device life.

• Network cable shall be labelled on both side with source as well as destination for quick identification during troubleshooting any kind of outage or in case any cable is accidentally unplugged.

• Device shall be free from dust at front, back and even top side. Dust at front or rear grills may block air flow leading to device heating. Dust layer at top may obstruct release of heat from top metallic surface. 

• Racks and Data Center hall should be free from dust to avoid devices sucking them with cold air flow and getting deposited at front grill and on components inside device. These layer of dust may absorb humidity from air and slowly leading to degradation of circuitry. Use of sticky mat or shoe cover is recommended to avoid dust coming inside from out.

• Devices shall be labelled properly to identify the device in rack. We can use device host name or device model serial number. Serial number of all devices may not be visible if devices are mounted on top of other as their serial numbers may be on top of bottom side. No IP address shall be displayed on any device to ensure information security as multiple vendor may work at different times inside Data Center.

• Ensure all dual socket device are powered from different PDUs in racks, not from same PDU. Any single socket device shall be preferably powered through ATS. This ensure power level redundancy is maintained for all device. Monitoring ATS and PDUs via a DCIM tool give added advantage to let engineers know when there is any defect in power of any PDU or ATS and help to know rack load trend for capacity planning.

Data Center Security & Infrastructure Protection:

• Ensure we have dual or multi factor authentication while entering Data Center and is working properly. It can be RFID card reader with PIN pad or with finger print/retina scanner etc. Anti Pass Back shall enabled in Data Center entry & exit to avoid tailgating and shall be tested on monthly basis to ensure its functional. EBG button shall also be functional to ensure door is release in case of any emergency. Change of PIN on monthly basis is good practice if pin pad readers are deployed for entry.

• Check and ensure that DOTL (Door Open Too Long) alarm activates if door is kept opened more than predefined time. Normally its 30 seconds. This is applicable for entry gate and emergency exit door, both.

• CCTV shall be covering all areas of Data Center hall including hot and cold aisle  and no blind spot shall be left. CCTV focus and coverage shall be tested every month to ensure all cameras are working properly, they are properly focused and their angle is not changed during any maintenance activity carried out inside Data Center hall.

• Functionality of Smoke Detector, Gas Based Fire Suppression System (GBFSS), Very Early Smoke Detection Apparatus (VESDA), Water Leakage Detection System (WLDS) and Rodent Repellent System (RRS) shall be tested every month and any issue identified shall be recorded and resolved. Test of fire suppression system can be clubbed with fire drill and training of all staffs involved in Data Center operation, maintenance and physical security.

• Dual UPS power shall be provisioned from two UPS sources till every rack via different power paths and preventive maintenance of each source and path shall be currently tested at least once or twice a year.

• Rack front temperature shall be between maintained at 23±1 degree Celsius and Relative Humidity(RH) shall be maintained at 50±5%. This ensure devices are properly cooled. Higher RH may lead to condensation of water droplets inside device leading to corrosion of circuitry and lower RH may lead to increased accumulation of static charge which may damage delicate IT equipment, also lower RH may lead to discomfort to engineers working for long hours inside Data Center like dry eyes, dried throat etc.

• Any visitor entry inside Data Center shall be reviewed and approved by Data Center manager for specific purpose. Visitors identity shall be properly verified by security guard before allowing entering inside Data Center premises and hall. Visitors entry shall be recorded on visitor register recording entry and exit time. List of tools and equipment carried by visitor inside Data Center shall also be approved and verified based on type of work to be performed by him/her. 

• List of authorized user having access to Data Center shall be reviewed on monthly basis to identify whose access shall be revoked and whose to be continued. Access of resigned resources shall be revoked ASAP.

These controls ensure your Data Center is secured and is maintained in a way for longer device life and Data Center up time. Also, you are ready for Data Centers audits like Infosec audit, SOC1/2/3, Tier certification, ISO 50001 (Energy Management System), PCI DSS Audit etc.

Go to Content List