In the wake of the increasing threats to the data center via cybercrime and the recent natural disasters occurring across the globe, AFCOM's Data Center Institute believes it is time to review the adequacy of existing data center business continuity and disaster recovery plans.
They maintain that the cloud offers new and dynamic approaches to the art of disaster recovery. It builds upon the techniques developed with load balancing, virtualization and geographical separation of assets. If done right, the cloud can drive multi-purpose utilization of assets, which makes disaster recovery more affordable and can provide significant ROI.
Leveraging the cloud for use in a disaster will require pre-planning and monitoring of the resources available in the cloud to ensure that what‘s needed is there when it is needed. This slideshow features eight key steps, identified by AFCOM, for taking advantage of the cloud for disaster recovery.
Click through for eight key steps needed to take advantage of cloud-based disaster recovery, as identified by AFCOM's Data Center Institute.
Perhaps the most important component in running DR within the cloud is to assign a leader to design and manage the environment. The complexities associated with building and managing this environment are deep and overarching. Coordination between application owners and data center managers must be planned and executed with precision to guarantee success. A strong leader with the authority to set the direction will be an imperative.
Standardization of the platform is important in reaping the benefits of the cloud. Having standard hardware, processors, OS versions and builds make transporting virtualized containers easier and more reliable. Choosing a platform is something that must be done in partnership with both the application architects and the data center managers. Having a mismatch will result in applications not porting into the cloud during critical times.
Application prioritization is traditionally necessary to determine which applications get first access to available CPU cycles during the recovery process. In the cloud, resources are infinitely available, but you must still prioritize the order of application deployment to ensure the most critical systems are up and running first.
Using traditional methods of data replication comes with a list of issues such as latency, mean-time to restore data and offsite storage. In the event of a disaster, your data is only as current as the last backup. Tapes must be shipped from the offsite storage location and then restored at the DR site. This can take days or weeks depending on location and disaster preparedness.
With cloud computing, new products are coming on the market every day to help create mirrors of data that are dispersed across cloud domains. This mirroring in effect protects the data from any single point of failure and removes the need to replicate data in a static fashion. As application traffic is transferred to other areas in the cloud the data is there to support them with limited disruption.
Network traffic routing and capacity in the cloud can be confusing and complex as cloud providers try to make it abstract and simple to manage. Some providers even claim that the network is invisible. The most important aspect is to make sure that the cloud provider, whether private or public, has enough bandwidth to support all of its customers. In catastrophic events such as 9/11, many DR facilities found themselves short of network capacity since they oversold their service.
Monitor your network usage across the cloud and see what levels are reached during peak times. Add all of the usage together across the cloud and determine if your hosting locations can accommodate the increased usage if any portion of the cloud goes out of service. Some providers will allow you to pay more to be placed in higher priority slots (above other customers). Make certain you address these issues before committing. Another trend to investigate is cloud balancing, where you are leveraging multiple cloud providers to keep your applications highly available.
Monitoring and oversight of application architectures will make certain you have what you need in the event of a disaster. If you are using the cloud as a DR site for hosting your application, ensure that all application changes are implemented in your cloud provider accounts. Any changes to the hardware or OS in one private cloud facility must be planned and replicated at the disaster site. This is a key advantage over traditional DR facilities where they may sit idle and quickly become outdated and unable to support the applications. New products today simplify how to electronically replicate your application environment in the cloud, across domains.
Implementing public, private or hybrid clouds raises many questions and security concerns. Can you trust the security of your data authentication, access control, encryption and monitoring? Protecting the data itself and/or access to the database are both areas of concern. Several methods exist including standard encryption tools and tokenization. Investigate the best solutions available by discussing it with your database tool provider. Whatever solution you choose, be aware that application performance can be impacted so make sure to balance risk against performance.
Lastly you need to understand where the data is ultimately stored. Several countries have governmental regulations that do not allow certain data to leave the country. Make certain that the data is categorized and assessed against regional data protection policies. Some cloud providers offer selectable geographical zones for CPU and data location.
