How to Design a Cloud-Ready Disaster Recovery Infrastructure
Are you in the cloud? Are your ready for a possible outage? Learn how to architect a cloud-ready disaster recovery and business continuity (DRBC) infrastructure.
March 28, 2014
Emergencies happen, entire platforms go down, but the business process must go on! Right? This recent DCK story says it best: “The Year in Downtime: The Top 10 Outages of 2013.” We’ve seen all sorts of issues happen with both cloud and private environments.
It’s important to create an environment capable of supporting business continuity needs. This means understanding the fact that the cloud can and will potentially go down. For example, in a recent major cloud outage, a simple SSL certificate was allowed to expire. This then created a global, cascading failure taking down numerous vital public cloud components. Who was the provider? Microsoft Azure. Over the past few years, organizations found it challenging to enter the DR and even business continuity conversation.
First thing’s first: it’s important to understand that disaster recovery and business continuity are two different business objectives. However, they can certainly overlap. Even today there is still some confusion about what the cloud can really deliver. It’s not so much how the cloud works, but more so around DR capabilities.
Whether you’re a small shop or a large enterprise, if you have haven’t look at the cloud as a DR or business continuity option, it’s time that you do. Already, organizations of all sizes are looking at ways to expand their platform – intelligently – into the cloud. The idea is to create a powerful, pay-as-you-go, recovery solution. For example, Bluelock already has some pretty powerful recovery-as-a-service (RaaS) options. Bluelock’s Recovery-as-a-Service (RaaS) solutions enable organizations to recover their IT resources efficiently and effectively when an adverse situation strikes, protecting you from loss of revenue, data, or reputation. The idea is to create a solution by combining the benefits of a multi-tenant cloud with the latest recovery technology to get a highly reliable, testable and enterprise-grade recovery solution at a compelling price point.
That said, there are specific cloud technologies which have become driving factors for better business IT redundancy. Before we get into the execution part of a DRBC strategy, it’s important to look at the solutions that help make it all happen.
Cloud-based traffic management. This is all about global traffic management (GTM) and global server load balancing (GSLB). Our ability to logically control traffic over vast distances has allowed the cloud to become a true hub for many different kinds of DR strategies.
Software-defined everything! We now have hardware abstraction at its greatest. Let’s face it, working with today’s modern networking controllers is pretty nice. We can now create thousands of virtual connection points from one physical switch port. Plus, we’re able to control layer 4-7 traffic at a much more granular level. Furthermore, solutions around storage, compute and security can further create powerful cloud resiliency models.
Virtualization and automation. Working with agile technologies like virtualization allows the sharing, replication, and backup of various kinds of images and workloads. These images can then span global data centers depending on the needs of the organization. Couple that with automation technologies and you can now span your platform from one infrastructure to another very seamlessly. Plus, you can better control resources which span numerous nodes.
New types of storage and hardware intelligence. Cloud-based storage solutions have come a long way. Whether we’re working with an open-source cloud storage platform or something more traditional like an EMC or NetApp, our ability to control storage is pretty advanced. Data deduplication, controller multi-tenancy, and fast site-to-site replication make cloud storage systems a powerful part of the DR process. Hardware solutions aside, the logical layer of storage is pretty powerful as well. Software-defined storage completely abstracts physical components and allows for VMs, data points and other infrastructure resources to scale very dynamically.
Now, let’s look at some ways that DRBC can really happen in the modern enterprise.
Here’s an important point: Organizations will need to conduct a business impact analysis (BIA) to establish effective recovery time objectives (RTOs) and the internal DR strategy.
Once that’s complete administrators can look at various cloud models to help them delivery true infrastructure high availability. From the cloud’s perspective, the options are numerous:
Hot site or active/active configurations
Requirements: Minimal downtime should be expected should an emergency event occur.
Note: This can be pricey. Basically, you would need a hot site that is always operational – sill, this may be a necessity.
Warm or cold site active/passive
Requirements: Depending on your BIA - Some downtime is allowed. But not prolonged.
Note: This is less expensive and can be adopted under a more dynamic public cloud model.
Workload-based disaster recovery or business continuity
Requirements: Not the entire infrastructure needs to be recovered – only certain services, applications, or databases.
Note: While still stored in the cloud – applications, databases, or other services can be either mirrored live or be provided in case of an emergency.
Backup-based recovery
Requirements: Downtime is not a major factor – but the application, or workload is still very important and needs to be brought up quickly.
Note: Similar to workload-based recovery, these cloud services replicate data, applications or other services to a cold VM-based backup. Should the need arise, either specific data or an entire workload can be recovered. Depending on the contract – this process is a bit slower.
Recovery-as-a-Service (RaaS or DRaaS)
Requirements: The beauty of this type of solution is the flexibility. You can work around an environment that requires no downtime – or one that can allow for a bit of an outage. Or, you can specify specific workloads or applications recovery should an emergency occur.
Note: Already there is integration into underlying virtual systems to creating a dynamic solution is not as much of a challenge. Plus, this is a powerful pay-as-you-go model that only charges you for what you recover; instead of just charging your for resource usage.
Remember, there are numerous different kinds of services being offered which are originating from cloud vendors and providers. The really interesting part here is that these products and services are being delivered to a larger vertical and more organizations of various sizes. With more resilience, better replication, and a lot more support, the cloud can be a pretty powerful platform for your DR strategy.
About the Author
You May Also Like