No one likes talking about Disaster Recovery – except maybe the occasionally entertaining brainstorming meetings where everyone discusses the risks – hurricane, flood, meteor strike, zombie apocalypse.
Risk discussions are followed up by defining DR tiers– instantaneous, minutes, hours, days – maybe even with metals used to define the various tiers, such as platinum, gold, silver, and tin.
Next up are discussions around what systems fall into each tier. Really those aren’t discussions – they are arguments between IT teams and with various organizations in the business. The arguments get more heated when the discussion leads to who is responsible for paying for all that equipment, software, services, etc., for supporting DR.
Finally, the periodic test – often set up for failure from the beginning as technical change management processes don’t account for DR systems in their processes.
Exhausted? These are just a few of the reasons why most companies don’t execute DR effectively – and honestly hope a disaster will never come along and truly test those systems.
Where to Start?
Where do we recommend Disaster Recovery planning start? Simple – a high quality, high tier data center in the low-risk location as the primary data center(s). That’s right – the core data center(s) where your equipment resides is the starting point for DR discussions. Inadequate power protection, high-risk locations, and insufficient redundancies all contribute to the potential for a disaster – and definitely increase the risk of more frequent outages that negatively impact the business.
Tier 3* data centers are the minimum requirement for a data center’s sufficient base availability for most organizations. For companies that seek greater peace of mind or require 100% availability, Tier 4 or Tier 5 data centers are required. Most companies cannot afford to build, much less maintain the redundancies required for Tier 3 and settle for Tier 2. Tier 2 data centers don’t allow for online maintenance of critical data center systems and don’t have necessary redundancies in case of critical data center component failure. All of the redundancies in higher-tiered data centers increase survivability in ordinary and more hazardous times.
* Insiders Tip – While researching data center facilities, do not assume Tier 3 is cheaper than Tier 4 or Tier 5. Many factors go into the costs of building and maintaining a data center, and often the highest quality can be delivered at the lowest cost.
Just like buying a house, the location of the data center matters. It matters a lot. Picking a low-risk location – one with the smallest risk profile to natural disasters – is critical in a data center location. Some are simple – California and Florida are easy examples with higher risks of earthquakes and hurricanes. Louisiana and the Texas Gulf Coast are two other easy locations to avoid. Some Midwest areas are of high risk, also due to flooding. Large cities, such as New York and Washington DC, have a perceived increased risk for terrorist attacks based on historical events. The United States Geographical Survey (USGS) publishes risk profiles based on various factors. Selecting a location based on the USGS risk profile can significantly reduce the likelihood of disaster.
** Insiders Tip #2 – Most applications will work just as well in a facility even a few hundred miles away, and some applications can be configured to be geographically independent of the user community. Don’t just research data centers in the local area – expand the research to the entire region. Many companies miss out on the best option by only looking or thinking locally.
Don’t Forget Reliable Power Delivery
Along with location, reliable and redundant power systems support the overall availability of the data center. Many corporate-owned and managed data centers have limited control over the power delivery and redundancy available. Function-built data centers – especially those at Tier 3 and above – have redundancy built-in from the building’s core installation. Higher quality data centers will determine their location-based power source and reliability.
An example is the Terminal Commerce Building in Philadelphia, PA – more commonly known as 401 N Broad Street – manufactured tanks during WW2. Dual power was implemented into that facility due to the war effort’s criticality – and converted into one of the most famous data centers on the East Coast.
Redundant Data Centers and Cloud
Most data center companies have multiple locations with direct communication links between those locations. Utilizing one for primary and another as secondary for either system redundancy or full DR can be a highly effective and reasonably economical strategy. Additionally, high-quality data center companies have high speed and redundant connectivity to modern cloud offerings that include DR services for hosted equipment.
Is DR required if hosted in Tier 4 or Tier 5 Data Center?
That is up to each company’s risk profile. Most boards require some level of DR though the IT organization defines specifics strategies. With a Tier 4-5 data center supporting primary hosted systems, simplifying the DR plan is a more viable option than lower-quality facilities.