Why does a disaster recovery plan matter today?
A disaster recovery plan has probably never been more important. As digital transformation accelerates the speed at which we all work, acceptance of lost connectivity or poor application performance is at an all-time low. Your disaster recovery plan ensures that in the event that an asset or connection failure, an alternative can be made available – quickly and efficiently.
The three parts of a disaster recovery plan
Let’s consider a service provider such as a streaming service or a business with a customer-facing application. A storage drive holds content while a server manages your membership and account details. Residing within a data center or edge computing location, a power distribution component receives 20kV, and transforms it to 230V to power the server and storage drive. At the same time, the room in which this happens is managed to a specific temperature to maximise performance and minimise the risk of outages. Finally come the network devices that carry the content to the end user, which in turn are reliant on optimal fibre to complete the transmission.
From this example, we can see that to achieve a seamless disaster recovery plan, you must protect three key elements of your infrastructure:
- The servers and storage devices that host your data and applications
- The power and cooling provided by the data center that houses the server
- The telecommunications network that connects the user with the server
Delivering your disaster recovery plan
These three elements provide the basis for best practice in disaster recovery. For example, key actions that most enterprises today will undertake include:
At the server level:
- The use of multiple storage devices that can be chosen dynamically, depending on their proximity to the user and the workload
- A domain name server (DNS) to enable redirection of the connection between user and an alternative server if the normal one is not working
- Virtualisation of the server, ensuring the application or service can effectively jump from one to another if the first fails down
- As high-bandwidth availability has become common in many data centers, it has become easier to develop architectures composed of geographically dispersed servers which provide multiple backup options
At the data center level:
- While the data centre architecture cannot be virtualised, power transformers, genset and UPS can all be designed with significant redundancy – so if one fails, another can instantly take 100% of the workload
- Most organisations will apply the same approach for their cooling systems
- Typically this redundancy will be 2N+1 – in which a fully paralleled backup system is available, further supported by additional components.
At the connectivity level:
- The most obvious connectivity solution is to allow diverse routing, providing two paths from the server to the point of consumption
- In some use cases it will not be possible to achieve this on an end-to-end basis. If a user is served by a single fibre, it results in a single point of failure
- In such cases the strategy is to use dual fibre access with diverse routing, allowing data to be pushed through an alternate route in the event of a failure or excessive traffic
How DATA4 supports your disaster recovery plan
It is clear that in any disaster recovery plan, the quality of your power, cooling and connectivity are critical. DATA4 is committed to providing its clients with an exceptionally high power quality – having experienced virtually zero power interruption throughout our history in operating data center campuses.
Our facilities are also carrier neutral, with a multiplicity of operators connecting to our data centers. This supports even the most demanding of customer use cases. Through this approach our clients benefit from the redundancy they need, and have access to the facilities required to build geographically dispersed infrastructures.