Cloud commuting or cloud computing—as it is popularly called—must necessarily take cognisance of Virtualisation. Today, every organisation has virtual machines that need to be backed up and recovered. But, how are these virtual machines different from physical machines? A basic understanding of the concepts will go a long way towards easing the process of backing up or recovering organisation data to or from the cloud.
What is Virtualisation?
It is true that virtualisation has been around for many years now. It was first used in 1960’s and remained a feature of mainframes for a long time. It is only in the last few years–with the emergence of desktop virtualisation, network virtualisation, and later the cloud—that virtualisation has entered mainstream computing. A close look at what the term means will reveal that virtualisation is an umbrella concept. It is used in different contexts to refer to different things in a production environment.
Desktop virtualisation is the process of running multiple virtual desktops on a single PC. Each virtual desktop will contain its own operating system and applications and can be configured to load up on appropriate user authentication. Application virtualisation is the process of separating the application from the operating system and running it within a “sandbox” that can reside on a remote server or stream in from a remote server.
Server virtualisation, as mentioned earlier, allows many virtual machines to run on a single server. The different virtual machines share CPU, memory, storage and networking resources. Storage virtualisation is the process of grouping together physical storage using software, so that all the storage devices appear as if they were a single storage drive. It must be noted here that virtual storage is not the same as the virtual machine. The virtual machine is a set of files that run in the server box to create an operating environment, while virtual storage runs in memory on the storage controller using software to create the impression of a single storage repository. All the files in storage may or may not be present in the single server.
Cloud vendors exploit these two virtualisation features to make their services scalable and redundant. Network virtualisation is the process of using software to perform network functions. This is achieved by decoupling the virtual networks from underlying network hardware and using software switches to move data around.
Not being prepared for disaster is not an option. Total environmental duplication is expensive. Attempting to recover only when there has been an incident is tempting the fates. These traditional approaches to disaster recovery create a lot of uncertainty and even stress for the organisation.
The growing sophistication of the cloud holds out a promise for backup and recovery of both physical and virtual machines. Cloud backup vendors are “disaster aware” and make elaborate provisions for recovery of their customer data. The disaster recovery efforts indirectly benefit customers who sign up for cloud based backup and recovery accounts.
At the core of the disaster recovery best practices that is instituted by the cloud vendor is—data replication. Replication helps the vendor rotate the backed up customer data offsite. The replications use high-speed connections for streaming the data and double protecting it. The replication servers are separated geographically from each other, to protect the information against natural disasters that may impact any single data centre. The vendor may create a hot site and a disaster recovery site with failover provisioning to ensure that customers always have access to their data even when the primary server experiences a shut down.
However, it should be noted that virtual machine replication is a complex task and not all cloud vendors have the technology required for the purpose. Traditionally, the data contained in the virtual machine has streamed through the primary server before it is backed up to the replication server. When the virtual machine has to be restored, the virtual machine can be brought online by restoring the virtual disk to the production stage. All this may take a lot of time, and time is the scarce commodity during a disaster.
Improved cloud replication management technologies allow cloud vendors directly stream the data from the virtual machine to both the primary and secondary servers directly. This enables instant recovery of the virtual machine (which exists in the production stage on the secondary server) even when the primary server experiences an outage. Disaster recovery plans can be implemented instantly. Recovery Time Objectives (RTO) and Recovery Point Objectives (RPO) can be tightly coordinated. Check with your cloud vendor on this point.
The integration of VM replication with the data protection solutions offered by your vendor is a bonus. It will empower your organisation and keep you prepared for all kinds of disasters.