Mechanism for Cloud Disaster Recovery

These are additional requirements that must be met for DR to be provisioned as a cloud service. These include network reconfiguration, security and isolation, virtual machine migration and cloning:

Network Reconfiguration

Cloud platforms allow a flexible reconfiguration of the network setup for an application after it is restored online in the backup site to ensure business continuity.

Two possible network reconfiguration processes to achieve this include modification of Domain Network Service (DNS)  or updating routes to redirect traffic to the failover site, especially for public internet facing applications.

However, a strong collaboration between cloud platforms and network service providers will make this a success.

Security and Isolation

Cloud platforms must provide leverage, mechanism and guaranteed assurances that potential disasters will not impact the performance of applications running in the cloud.

Security concern to be addressed include the privacy of storage, network, and the virtual machine resources being used.

Likewise, clouds must guarantee that the performance of applications running in the cloud will not be impacted by disasters affecting other businesses.

Virtual Machine Migration and Cloning

Although not currently supported, cloud platforms are expected to allow VM migration in or out of the cloud. This can be made possible as cloud exposes additional hypervisor level functionality to customers while optimizing migration techniques for the WAN environment.

Virtual machine migration and cloning is very important for the following two reasons:

  • Provides simplified failback procedure for moving an application back to its original site after a disaster has been managed.
  • It can be used to improve planned maintenance downtime.

Disaster Recovery as a Service

Disaster recovery as a service is perceived as a future low-cost service designed after the cloud computing nomenclature is expected to offer flexible replication be it physically or virtually.

This phenomenon allows recovery of data center infrastructure and critical servers to be replicated in the cloud as a service.

The architecture is configured with pre-built options to support virtual recovery environments characterized by network connectivity, security and server failover.

In the event of a disaster, provisioned backup and applications can be executed on DR services until the backup to the primary site is obtained.

The table below presents a comparative analysis of both traditional and service-oriented disaster recovery approaches.

Traditional Disaster Recovery Disaster Recovery as a Service
This requires a secondary physical DR site and thus translates into further costs of manpower requirements, site maintenance, and operational costs, additional data center space, connectivity and servers. This only makes a virtual machine replica of physical or virtual servers available at the primary data center. Additional cost attraction is only for storing the replica and application data in a suspended mode as well as data replication from primary to secondary site for data synchronization.
If a disaster occurs, virtual machines are brought online to substitute for the primary site.
A physical DR site is operated when an actual disaster occurs. It takes more time to bring the DR site live leading to huge data loss. The DR site can be brought online within minutes after the event of a disaster translating to data loss of a little timeframe.
It requires manual connection if the connectivity is unavailable to commence site operations. It is triggered seamlessly regardless of geographical location using the internet.