BY DAN KURTZ, Vice President of new business development for LightEdge Solutions
Datacenters are designed to have back-ups in place so they do not fail on the physical level. The same should be true for the Clouds in which they are storing data.
Cloud-based resiliency ideally reduces the risk of failure because critical information is spread across several datacenters using networks, data replication and traffic switching. This approach will become even more widely used and therefore, more disruptive, according to 451 Research, an information technology research and advisory company that analyzes emerging technology.
The level at which Cloud-based resiliency is adopted will vary. Businesses will adopt different resiliency strategies for different situations, applications and internal requirements. Public Cloud operators Google, Facebook, IBM and AWS use cloud-base resiliency, though some use a hybrid approach.
New ways of storing data in preparation for disaster recovery may mean replication, global traffic management, data distribution and other methods are now being used. In the event of a failure, resiliency could mean a more cost-effective way to hand over the information to another venue that is supported by a reliable network.
There are still many unknowns about the public Cloud, and distributing data could have negative effects on whom builds and designs datacenters and the market for datacenter equipment and services.
There are also other factors to consider when switching to a Cloud-base resiliency approach:
- There could be more costs because ideally there are at least three separate datacenters involved.
- The physical infrastructure of the data center could be protected at a lower level because it’s cheaper to purchase the information technology equipment to spread out the data.
- Some applications are not suited for this model and would need to be supported with traditional models or converted.
- More complex information technology or management software will be needed to keep track of things. Anytime something becomes more complicated, it runs the risk of failure.
- Companies will need to ensure they have the IT skills to implement this type of system.
- This system requires more trust in software and third-party providers, which could mean a lack of transparency and a loss of control.
- The network in between datacenters will become more critical, and there could be questions or issues with performance and capacity.
The use of Cloud-based resiliency effects suppliers, service providers and operators more so than other types of resiliency. For example, this type of resiliency must operate at scale to ensure other data centers have the capacity to pick up the load or partial loads of other ones. Many public Cloud providers will organize their resiliency efforts in zones and regions to ensure there is sufficient capacity in place, including in the network, if loads need to be shifted from place to place.
That’s why the following must be taken into account, if data is going to be spread out among multiple datacenters:
- At least three datacenters need to be involved to ensure the network has sufficient capacity and variety.
- The need for high levels of redundancy in power, cooling and information technology at each site may be reduced because loads are running at multiple datacenters or traffic can be switched to any datacenter.
- The appropriate infrastructure must be in place to make sure applications can run at any site or across all at once.
A disaster recovery plan may need to be redefined because it no longer requires the same processes and resources.