Moving from backup to availability

CIOs must free their organisations from complex backup strategies in order for storage and virtualisation systems to protect against failures and cyber attacks.

Modern enterprises are managing a complex network of clients and servers, often without a concerted approach to availability. Adding to this complexity is a growing mix of mobile devices and cloud services where data is generated outside the reach of routine backups.

Traditional backup and recovery poses a challenge, in that the process is prone to lengthy delays and even data loss due to errors and lack of testing. In the case of an online system, recovering from a point-in-time backup can directly impact the performance of that system and the organisation’s ability to service customers and generate revenue.

More IT-driven customer interaction processes call for more data and application availability, not just backup. A modern availability strategy enables companies to re-architect their traditional backup processes for faster recovery times and closer recovery points, without impacting production workloads, apps, virtual machines or data centres.

Best-of-breed approach

There is a real change in how enterprises are managing their infrastructure, with physical servers being virtualised and collections of virtual servers being orchestrated into private clouds. Telsyte research indicates 80% of Australian enterprises are using or investigating server virtualisation technology, making Australia one of the more mature markets globally.

This trend towards more server and storage virtualisation has profound implications for backup and availability. A strategy to keep businesses always on must now include the capability to perform availability for both physical and virtual machines.

Along with the growing trend of virtualisation is the adoption of more than one type of hypervisor. VMware is the leading global supplier of enterprise virtualisation technology; however, in recent years Microsoft’s Hyper-V virtualisation platform has risen to prominence as an alternative. Today, more than a third of enterprises using virtualisation are using more than one type of hypervisor, according to Telsyte’s research.

Given that no vendor dominates the virtualisation and private cloud space, a best-of-breed approach to availability is essential for the best outcome across multiple platforms. Using the right backup and availability solution for a virtual environment can reduce complexity. For example, some backup solutions require agents which can inhibit the benefits of virtualisation, and add to the management complexity and cost that virtualisation aims to reduce in the first place. An availability solution is often the right fit for today’s hypervisors, rather than agent-based backup applications or systems.

In-house, cloud or hybrid

With more infrastructure going virtual, Australian organisations are well equipped to take advantage of infrastructure-as-a-service (IaaS) from local and international cloud providers.

The next step on the path to an on-demand infrastructure is orchestrating virtual machines into a private cloud. Telsyte research indicates that the penetration of private clouds among Australian enterprises will approach 45% by 2019. This is happening alongside strong adoption of public cloud IaaS, which is forecast to reach 85% penetration by the same year.

There is a perception that using a cloud service automatically means data is protected and highly available. This is not true, and workloads both on- and off-premises can fall victim to poor data management and downtime.

There are numerous options for protecting data to and from the cloud, including backup-as-a-service, disaster recovery and replication-as-a-service; however, the solution must also be fit for purpose in the cloud, as many services do not afford the equivalent level of features and access to the hypervisor.

With a mix of private and public clouds in use, the future is a hybrid cloud environment where the right workload is hosted on the right infrastructure. Working towards a hybrid architecture will bring a new level of flexibility around how and where applications are hosted and what elasticity organisations have during peak load times.

Tested data protection

With so much time and resources spent on backup and disaster recovery, it is crucial organisations understand the effectiveness of these activities. It is one thing to perform a backup or disaster recovery action, another thing to verify the success and efficiency of that action. If the backup is not successful then the investment is wasted.

To determine effectiveness of data protection processes, CIOs must investigate how the organisation’s backup systems are performing and the frequency of testing and restoration. The backup medium itself will play a part in how well the data is protected. Online, disk-based backups can be tested regularly, whereas backups to tape can be left untouched for long periods of time. In the event of data needing to be restored, recovery from tape can be time-consuming and error-prone if the media has decayed.

To eliminate the time it takes to recover data in the event of a problem, organisations should minimise the time it takes not only to perform backups but to test them as well. Combine this with a low recovery time and the business is on the right path to availability.

It is crucial to ensure all backups are tested. Backups are not infallible, and it is a misconception to believe that once data has been backed up it is safe for restoration. This ability to ‘test as you go’ is pertinent for disaster recovery — another area which can be costly, time-consuming and difficult to test. And it is just as important to test disaster recovery and remote backups as it is to test local backups.

In conclusion, CIOs must work towards streamlining disaster recovery processes to avoid dependency on manual intervention and potential loss of revenue if a disaster strikes and an untested architecture fails to take over. Today’s availability solutions offer more than simple recovery operations. The good news for CIOs is there are now more local options — including local infrastructure of international providers — for disaster recovery in the cloud, which can augment existing data centre operations.

Image courtesy Ervins Strauhmanis under CC BY 2.0