Companies have to protect their business critical applications. Avoiding data loss and downtime for the mission critical applications is extremely important.
Consider a scenario in which you have been informed that a storm is coming and will hit your datacenter.
If all the application workload can be moved to another company’s data center, your operations can proceed normally.
Moving a load to another data center can be seamless and I will be covering this seamless operation under active-active data center design articles series in orhanergun.net
Below chart illustrates major disaster causes in general but percentage might change based on your location,region , infrastructure of a city and so on.
In order to avoid disaster three critical elements play a key role.
- We need to be informed that disaster is coming.If it is mandatory maintenance in the data center and data center is operated by the provider, Service Provider should inform the customers. If it is storm or flood, we should be informed by the Government entities.
2. If the target is to perform seamless migration for disaster avoidance, data canters should be close enough to perform live migration.
Vmware calls it Vmotion, Microsoft calls Hot Migration.
With live migration, idea is to copy the RAM , memory pages of the running virtual machine from one data center to another data center.
Specific replication technologies should be in place for the storage network, otherwise live migration might be impossible for the virtual machine which has large amount of data since the period that we have to move data is limited to couple hours in many cases.
Disaster avoidance is hard if not impossible with physical servers and without some sort of replication technology.
Last but not least, the network link carrying the applications workload between the datacenters.
If you need to carry only 500GB of traffic from one data center to another within 2 hours, although depends on which transport mechanism is in place ( Ethernet, Dark Fiber , Sonet/SDH, DWDM/Layer0 and so on ) you need to have at least 100Mbps link between two data centers.
Although disaster avoidance includes the term “disaster,” this process is used for maintenance purpose as well.
There are times you migrate a workload to another datacenter for scheduled maintenance.
Many other technologies should included to support a seamless user experience during disaster avoidance such as replication,IP localisation, NAT, clustering, inbound traffic optimisation, session synchronisation and load balancing are just a few of them.
What about you ?
What is common disaster scenario in your region ?
Do you have a disaster avoidance strategy?
Let’s discuss in the comment box below.