Lessons from Hurricane Harvey: Ensure Your Data Center Can Weather Any Storm
Between 9 and 15 named storms are predicted this year, but there are steps that data centers can take to ensure they stay running during catastrophic weather events.
Netrality Wins Against Hurricane Harvey
Hurricane season, which lasts from June to December, is upon us. This year, the National Oceanic and Atmospheric Administration predicts we’ll have between nine and 15 named storms, with four to eight of them becoming actual hurricanes. Half of those turning into major hurricanes are predicted to have winds over 100 mph.
On the East Coast and near the Gulf of Mexico, hurricanes are one of the most serious threats to IT infrastructure. When systems go down as a result of their ferocity, it not only impedes business and government operations, but it also affects communications, equipment, and services that play a vital role in emergency response.
So, how do data centers ensure they can remain up and running during even the most catastrophic weather events?
To answer that question, let’s take a look at Netrality’s colocation data center at 1301 Fannin in Houston, which experienced zero downtime during Hurricane Harvey, one of the most devastating storms in American history. The facility proved itself a model of sustained operations for enterprise customers, even as Harvey unleashed over a trillion gallons of water onto the city.
1301 Fannin is located on the high side of the Central Business District, an ideal location for critical infrastructure during storms and other natural disasters.The building’s 200 mph-rated windows kept wind and precipitation out. Flood risk was also mitigated by the fact that 1301 Fannin is not connected to the underground downtown tunnel system, which often floods when hurricanes hit. In fact, the nearest flooding was nine blocks away.
To ensure the backup generators keep the building running when the power goes down, the facility is equipped with N+1 generators, N+1 Uninterruptible Power Supply (UPS) systems, and 70,000 gallons of fuel on site, all of which provide approximately two weeks of sustained power. (Just in case additional fuel is needed, Netrality has first-priority contracts with reliable providers in the area.)
Because the potable water system is powered by a series of backup generators, tenants of 1301 Fannin may occupy the building during the storm, per the City of Houston policy, while other buildings have to chain their doors closed. Finally, the data center also stocks weeks of ready-to-eat meals and provides all necessary amenities – including showers and bedding – for a prolonged period.
In other words, 1301 Fannin was prepared to withstand a historic storm, and it did.
What All Data Centers Should Do to Prepare for Hurricanes
So, what lessons might other data centers learn from 1301 Fannin?
First of all, prudent data center operators take emergency preparedness seriously, formulating detailed system configurations, conducting regular maintenance and robust testing, and developing comprehensive backup plans. With the proper preparations, as we’ve seen, data centers have the ability to remain operational even during the most severe weather events.
However, there are key preparations that both data center providers and tenants must take to ensure operability during a hurricane.
Data center providers must:
Design data centers appropriately
Ideally, before a data center is even built, it should be designed with hurricane preparedness in mind. This includes having a redundant physical layout with duplicate core operational components, such as fuel storage areas and generator spaces. Data centers should also feature flood-proof foundations and windows fortified against hurricane-level winds. If this is not the case with your current data center(s), it is all the more crucial to follow the remaining recommendations.
Test regularly and thoroughly
Regularly perform integrated systems tests, where power is completely shut down at the facility to test the effectiveness of battery backups and generators. Keep detailed logs of when the tests occurred and the test results.
Audit testing procedures
Audit testing procedures at least once a year to ensure backup power and cooling can maintain the data center’s current power load. You may have added more clients and servers since the last test and therefore have a greater footprint to manage. It would also be wise to re-evaluate your data center’s disaster recovery plan each year before hurricane season begins.
Tenants who connect at data centers must be sure to:
Follow the “3-2-1” rule
Make sure you follow the “3-2-1” rule: Have three copies of data, on at least two different types of media (tape, disk, cloud, etc.), with at least one copy located off-site.
Always have support ready
Make sure you have 24x7x365 support through a centralized operations center. This ensures all data center equipment and activity is constantly monitored, and that you will be alerted to all issues in real time.
Have a regional disaster recovery plan in place
In case your backup and redundant components fail, it’s crucial to have a regional disaster recovery plan in place with local support teams to provide immediate on-site support. Food, beds, and necessary supplies should be available at the facility to sustain an extended period of support.
Natural disasters present a huge risk for enterprises. Leveraging the dependability and infrastructure support of an interconnected data center is integral to keeping mission-critical applications in the cloud up and running at all times. 1301 Fannin’s redundancy capabilities and key electrical, mechanical and security offerings make the facility an ideal example of how a data center can support mission-critical applications and services for enterprises throughout any type of emergency or disaster.
To learn more about how Netrality’s interconnected data centers are best-in-class for disaster preparedness, contact us.