Disaster Recovery
Don’t make the mistake of justifying your lack of a Disaster Recovery (DR) plan by thinking, “Cyclones rarely visit my neck of the woods,” or “Floods occur only every one hundred years.” All of these statements may be true. However, disasters on smaller scales happen far more frequently, often hundreds of times more frequently, than the big ones.
Smaller disasters — such as building fires, burst pipes that flood office space, server crashes that result in corrupted data, extended power outages, severe electrical storms, and so on — occur with much greater regularity than big disasters. Any of these small events can potentially interrupt critical business processes for days. In time-critical, service-oriented businesses, this interruption can be a fatal blow.
When disaster strikes, businesses without DR plans have an extremely difficult road ahead. If the business has any highly time-sensitive critical business processes, that business is almost certain to fail. If a disaster hits a business without a DR plan, that business has very little chance of recovery. And it’s certainly too late to begin planning.
Businesses that do have DR plans may still have a difficult time when a disaster strikes. You may have to put in considerable effort to recover time-sensitive critical business functions. But if you have DR plan, you have a fighting chance at survival. MeshBlox Services can provide the experience and expertise needed in devising a plan to meet your needs.
A major component of a DR plan is the Data Centre Recovery Solution, which answers two major questions:
- Where will you go?
- Recovery site
- Internal alternative
- Commercial hot-site
- Hosted high availability (internal or commercial)
- Location
- Geographic separation
- How will you connect?
- Data connectivity alternatives
- Solution to include connectivity to your recovery site
- Multiple paths and pre-configured
- Voice considerations
The strategies to recover a data centre from a disaster are discussed in the following sections:
Drop Ship Strategy
Drop shipping complete server requirements for disaster recovery is a technique in which your organisation would identify in advance the selected hardware required to meet your recovery objectives. A third-party provider would supply the hardware and would ship the agreed-upon contents based on an agreed-upon Service Level Agreement. The recovery time for a drop-ship solution ranges from 24 hours to seven days plus. The expectation is when the equipment arrives, there is provisioned supporting infrastructure (computer or server room) in place to plug the solution into.
With the DataBlox series of containerised data centres, it is now possible to meet those expectations regardless of the site, by dropping in the supporting infrastructure in the form of a MeshBlox containerised data centre, along with a Cummins® power generator directly to the affected site.
Cold Site Strategy
A cold site is an empty, vacant facility with sufficient electrical services pre-wired, HVAC and racking that can be set up with equipment and accommodate personnel in a short time. The equipment and other resources on which the business depends will need to be procured, installed, and configured at the time of a disaster. Additional time exposure can be the length of time it takes to procure or move equipment into this facility. As a result, setting up the site will take longer (days to weeks) than setting up a site with the required equipment already installed and running (hours to days). To use a cold site, your Recovery Time Objective (RTO), which determines how long your organisation can afford to remain down without severely impacting business, must be long enough to allow you to acquire, configure, and install all necessary systems. While a cold site keeps your initial costs very low, the cost of shipping hardware and software immediately can be quite expensive during an actual disaster.
Generally, leasing space in a hosted data centre is the same price whether it is being used or not and can be an expensive option as a cold site. Building a spare server room in branch office requires a large capital investment and also fixes you to a location that may become sub-optimal. A MeshBlox containerised data centre becomes the best option as it can be moved to any location required, preconfigured ready to take equipment.
Hot Site Strategy
Hot sites comprise all of the required servers, communication infrastructure, and staff resources necessary to meet business requirements in the wake of a disaster. Some hot sites are fully equipped with office suites that also include desktop recovery, telephony, fax machines, and printing capabilities, all set up and waiting for you. The servers at the hot site are already powered up, software is pre-loaded, and the network infrastructure is already installed and connected to the supporting telecommunications carriers. In addition, the hardware and software at the hot site will be consistent with the installed inventory in use at the primary site. This supports a typical 12 to 48-hour RTO and a 24-hour Recovery Point Objective (RPO).
Server rooms operating within a commercial building are often teir-1 facilities, leaving critical services venerable to another disaster. A DataBlox can be configured as either a Teir-2 or Teir-3 data centre facility and can be deployed as the alternative to a primary site that is a traditional data centre facility.
Alternate Site Strategy
Consider a disaster in which the entire data centre or building is destroyed. In this scenario, everything that supports all critical business functions, current data, and server infrastructure is made unavailable. A duplicate site, owned and operated by the company, is an alternative. In other words, to ensure the successful recovery from a disaster, the organisation chooses to self-procure the solution. You can just store all your critical information, from backup tape media all the way up to a complete duplicate of complete hardware infrastructure, running warm or even with full replication. Alternate sites are also known as backup sites and are used in testing and in disaster recovery.
The DataBlox range begins with the little DataBlox 10, which is a 2 rack facility, complete with AC housed within a 10” shipping container, perfect for storing tape backups and for attached storage solutions. The DataBlox 20 provides up-to 7 racks capable of providing the space for a complete duplicate of hardware infrastructure.
Mirrored Site Strategy
A mirrored site is a fully redundant systems strategy. It supports organizations where traditional tape backup is not considered good enough to achieve the RTO and RPO goals for your most critical applications. Using a mirrored site involves maintaining a complete duplicate of all hardware identified as mission critical. The hardware is fully loaded with a complete, licensed copy of the operating system and all related licensed program products. In the event of a disaster, the organisation simply needs to shift operations to the alternate mirrored site and resume operations.
The site would have redundant network access that would permit users to connect to it with very little down-time. Therefore, the recovery timeframe is extremely short. All of this is accomplished by implementing a high-availability third-party solution, with the assurance that the solution is fully managed 24/7. If you promise to the business a fail-over solution in 45 minutes, for example, it must work. The integrity of the data and application access must be solid.
High availability will also provide for increased RPOs. At risk is only the last transaction. Latency considerations for transaction volume must not factor in any delay in processing the data. This implies owning and operating a big enough data communications connection between both facilities to ensure minimal or no latency of data. Lastly, implementing a mirrored site also involves some strategic thinking as to the location of the site. Geographic considerations are a must.
Which leads to the benefit of having a DataBlox; a mobile containerised data centre that can operate anywhere, not just where an existing data centre is built. In fact, the containerised data centre model goes a step further into a new genre of DR solutions, and this is through a Distributed Data Centre model. This model consists of many smaller data centres inter-networked in a mesh like manner to provide an incredibly high availability and high performance by allowing resources to be located close to end users lowering latency and increasing bandwidth.
Disaster Avoidance Strategy
With the advent of the MeshBlox containerised data centre, a new solution exists for Disaster Recovery. This is not an optimal solution, however in some circumstances it does make sense particularly from a financial point of view, and this is Disaster Avoidance.
In most occasions it is very hard to detect a problem before it occurs, but in the case of incoming large storms or political unrest potential disaster can be forecast. It is in these circumstances the most appropriate disaster recovery option is to move all personnel and equipment out of harm’s way. Because of the shipping container form factor, the DataBlox data centre can be transported away from the potential danger.
For more information on Disaster Recovery Planning or any of the strategies listed earlier, please contact a MeshBlox representative.
