Virtual servers: a new wave of IT technology that is taking the world by storm. The idea of virtual servers is a logical and natural development; everyday physical servers are usually assigned to only one task or application to optimize their performance, but they are capable of much more. Virtual servers capitalize on this latent potential by creating multiple servers within one physical casing; basically, the idea is to take one physical server and split it into multiple servers within itself that have the same processing power as the original. This obviously saves time, money, and space, and is naturally becoming the norm in computer networking. However, in the world of IT, new advancements usually come with new challenges: the main problems that networks with virtual servers experience are data backup related; specifically with backup performance, tracking, granularity, and media.
The first of a series of problems is backup performance, and it is encountered when data is being transferred from a set of virtual servers. In the past, the backup infrastructure would only be dealing with one server at a time, so when it is faced with multiple streams of information it can become overloaded. This issue is usually solved with specific apps, such as VMware’s vStorage APIs for Data Protection, that help manage the new format of data backup. The choice of media on which the data is stored can also pose a few problems. Most virtual server data is backed up using changed block tracking, and the data ends up on a disk. Regrettably, disk storage isn’t the cheapest option when dealing with data on a large scale, and taking it off site isn’t a choice as well. An alternative to consider is tape storage, as it allows for the creation of a full system synthetic backup.
Tracking becomes another issue of due to virtual servers usually not being backed up directly, but through a host hypervisor; keeping track of each virtual machine becomes complicated as they move between physical servers and storage. The solution is to use a datacentre object to abstract the reference to the VM, therefore allowing the backup process to no longer depend on a physical host.
When recovering data from a virtual server, it is often that there are only a few files that are specifically needed rather than the whole server; backup software sometimes will have trouble understanding the VM format and make finding a single file difficult, as it is used to a single physical server format. Users of virtual servers need to make that their backup software is up to date and capable of handling VM formats. Last but not least, since VM’s are stored as data on disk instead of being an actual, physical machine, snapshots are commonly used to ensure backup. Even this measure doesn’t guarantee full protection, as a fire, flood, or other emergency would cause total array failure – against which snapshots aren’t much use.
So, while virtual servers do come with a slew of benefits, one can plainly see that they demand a high level of maintenance and awareness when it comes to data protection and backup; multiple methods need to be put into use in order to ensure that data is safe.