The cost of downtime varies significantly across various industries. Estimates range from as much as $6,000,000 an hour for a brokerage business to $90,000 per hour for media services. Backup and disaster recovery or business continuance enable an IT organization to recover from the loss of a file, object, data set, application, or an entire data center.
The first step in ensuring business continuance is to have a second copy of the data as a minimum. Backup has evolved over the years with more sophisticated software and target hardware to enable faster and more frequent backups. These advances have increased availability, but backup remains a complex and expensive process, traditionally requiring data replication to a secondary data center. These copies are then leveraged by compute and network resources at the secondary site that are either dedicated or on standby to restart operations.
Cloud computing and storage are now enabling more cost-effective business continuity, with options ranging from utilizing a collection of point products and cloud services to a full disaster recovery as a service offering
Public cloud computing and storage continue to grow in adoption, with IT organizations and businesses attracted by the self-service, on demand, and elastic provisioning characteristics of the cloud. However, many IT organizations have concerns about potential control, security, and performance issues with a multitenant public cloud. A private cloud provides the scalability and self-service aspects of the public cloud, but with resources dedicated to the IT organization. This approach addresses the concerns with a multitenant environment. Most enterprises have a significant investment in on-premises infrastructure that is usable in the onsite private cloud deployment, which helps lower the cost of a private cloud.
Most IT organizations are opting for a hybrid approach to cloud. A hybrid cloud utilizes both a private cloud for more performance and security sensitive applications and a public cloud for both processing/storage overflow and less critical applications. Where a public cloud typically supports a “one size fits all” capability for performance, security, and compliance, a hybrid approach allocates either private or public cloud resources based on applications needs. Orchestration products assist with the interaction between the internal/private cloud resources and the public cloud.
One of the primary use cases in an enterprise utilizing a hybrid cloud is backup and disaster recovery. When data is backed up, the data center typically keeps one copy onsite to enable quick restoration in the event of a hardware failure, corrupted file, or accidental erasure. The public cloud is then used to store the second (and additional if required) copies of the data. A second copy is necessary for the event of the loss of the primary data center. The public cloud approach is far less costly than secondary data center that replicates the entire primary data center infrastructure. Multiple solutions are available to copy backup data to a public cloud. One effective approach is using an integrated backup appliance that serves as an initial onsite target for backup data. The appliance, which can be physical or virtual, then copies the backup data to the public cloud at intervals defined by user-created profiles.
A second approach is to store all primary data in the cloud, and then replicate copies to other cloud locations to provide an extra copy. An onsite appliance achieves higher performance for active data by caching active data locally.
For any of the cloud backup approaches, having a reliable way to continue business operations in the event the primary site is not accessible is imperative. The disaster recovery as a service (DRaaS) provider provides compute, networking, storage, and application accessibility to retain business continuance in the event of a primary site outage. The DRaaS provider defines the needed resources and the process for both the cutover to the disaster recovery process, as well as the eventual restoration of primary site services. The amount of time required to restart operations is dependent on how dedicated (and expensive) the DRaaS solution resources are. It is critical to define the DRaaS provider’s time commitment to restore operations, along with the ability to respond to a regional disaster that will put a strain on their resources.
Many other business continuity solutions include some derivation or combination of the solutions mentioned above. EMA is undertaking a research project to identify both the end user requirements and vendor/service provider opportunities for business continuance in the age of the cloud. If you are an end-user, customer, vendor, or service provider and have thoughts or questions regarding this upcoming research, or you wish to chat about the topic, please feel free to contact me at firstname.lastname@example.org or 303-543-9500 X 134.