Disaster Recovery and Business Continuity planning is one of the biggest single concerns to any data-dependant company today. Indeed, it has been one of the largest ongoing challenges for many years. One of the latest game-changers in this space is the growth of cloud based services. Certainly the hype around cloud is reaching fever pitch, but removing some of the layers of this hype off the top can reveal to us the real utility value in utilising cloud infrastructure services to meet Disaster Recovery requirements. At Hypernode, we’ve invested a lot of time (and money!) in breaking down the complexity to ensure that capable, commercially advantageous and objective based DR solutions can be accessible to everyone.
There are really two ways to look at the cloud in relation to DR and BC practices.
The first is to utilise cloud-based infrastructure to provide the DR layer to protect data stored on-premise, in public clouds or alternative datacenters. This is what we would consider the “Hybrid Disaster Recovery” model. The second idea is to take an entire on-premise data set (applications and data) off-premise into the datacenter and subsequently protect it within the datacenter offering. There are reasons for both approaches depending on the customer need that we can examine in more depth.
The Hybrid Disaster Recovery mode
The Hybrid Disaster Recovery model looks at methods of protecting, securing and retrieving access to your data in a disaster or data loss scenario by retaining backups of your information in multiple locations. As previously mentioned, this could be cases where your production data is stored on-premise on customer owned hardware right up to an existing public cloud platform such as Amazon or Azure.
Now, the case for providing datacenter based DR for on-premise systems may be more obvious – an on-premise server environment typically suffers a higher risk of loss of power, connectivity or physical equipment failure. Most internal server environments generally don’t have the kinds of facilities and systems redundancy we build into a full datacenter! Enacting a DR process to move production data offsite to a private cloud can be a great way to add appropriate layers of business continuity to your business.
But the second scenario, where the production data is stored in a public cloud or alternate private cloud environment may be less clear to you. After all, we did just state that datacenters have significant investment in redundancy to prevent loss. Therefore, should your public cloud data not be already safe enough? The answer to that question really depends on how important access to your data is. Let’s consider the case where you have production applications and data stored on a platform like EC2/S3, for example. These are indeed very redundant systems, but your ‘connectivity’ into these systems may be far less so. In cases where the routing between you and your cloud provider breaks down, your hands are tied. This is where having alternate DR in a separate facility, with different route paths and carriers, would enable you to enact a business continuity plan to continue working.
This is the same reason why Hypernode operates multiple facilities, including a dedicated facility in Central Coast NSW, as part of our DR product range – it allows us to deliver site, facility, location and route path redundancy as part of a single DR package. With any vendor providing you cloud services, be sure to break down your DR and BC plans and processes to ensure that there is enough end-to-end redundancy to fit your business requirements (in terms of Recovery Point and Recovery Time Objectives at a minimum).
Migrating straight to the Cloud
The second way for us to look at DR and the Cloud is to examine the advantages of migrating an existing on-premise application straight to the cloud. In the previous example, we examined the idea of backing up an on-premise system into the cloud while retaining production workload on premise.
There are various considerations we need to make in determining whether such a structure will work effectively – for example, does the connectivity between the premise and the cloud provide enough bandwidth to keep up a throughput rate that will meet the Recovery Point Objective. If not, can it be upgraded or replaced? Is it commercially viable to do so? We may find ultimately that backing up straight from on-prem to cloud is not ideal. In order to provide a better DR package, it may be more advantageous to move the production workload to the cloud instead, and then take advantage of DR services, or multi-site datacenter backups to meet the DR needs.
As you can already see, there are an extreme range of options around how the cloud can be used to meet DR requirements, or how DR requirements can be applied to cloud environments. It’s a bit of a minefield! Hypernode has extensive experience in building a variety of the solutions mentioned here. If you’d like to learn more or get some help in meeting your DR needs, give the Hypernode Data Protection Team a call.