Ref Link; cio.com blogs
Posted by: DBTK in Best Practices
High availability for full server protection has never been more important. As virtualization and blade technologies continue to rapidly be adopted, IT infrastructures are becoming smaller and more efficient. Doing more with less has been a centralized theme for IT managers this past year but “less” doesn’t mean ignoring backup and recovery strategies. If anything high availability solutions are also being consolidated, easier to deploy and much more cost effective. Most everyone has seen the availability curve describing the cost of a low recovery point (RPO) or time objective (RTO) using some medium like tape can be low cost but not effective to achieve high availability for business critical applications. As the RTO and RPO window shrinks typically the cost increases to achieve rapid five 9 availability. This is no longer the case and if anything the paradigm is shifting and becoming the exact opposite because the cost of downtime or the length of time it takes to recover is more expensive than preventing it.
The Backup Conundrum
A few years ago backup and recovery was relatively straight forward. IT managers had one server that ran one critical application, such as Microsoft Exchange, SQL Server or BlackBerry Enterprise server, and the process was to backup that one server and have the ability to recover in the event of a failure. Today, data centers have consolidated locations, servers and virtual infrastructure where one physical server may be running multiple applications and virtual machines that service an entire organization. Should that one server or power to that server rack fail, the impact is exponential because of this multi-functional architecture. In this scenario, tape will not be sufficient to restore operations as fast as required. So although a consolidated virtual infrastructure is much more efficient the cost of downtime has increased 10 fold from original ROI studies because the impact is greater.
Now add to the mix the combination of operating systems and virtual platforms. The high availability solution for Windows isn’t the same for Linux, or VMware vSphere to a simple physical server. Or is it?
Unifying Your High Availability Solution for all Platforms
The answer is yes! It is and can be this simple to use a unified solution to provide high availability as well as the same backup and recovery strategies to all combinations for infrastructure, cross platform as well as hardware. Some of the misconceptions of recovery are if a certain brand of server fails that an exact copy of that server is required in order to recover. This is also false; servers can be recovered to any server platform no matter what the hardware vendor is. This is referred to as hardware independent full server recovery. The process involves the backup and recovery of the system state of the server, applications as well as associated data. And some really smart engineers in Indianapolis, Indiana figured out how you can eliminate hardware dependencies and recovery cross hardware platforms. This same innovative technology can be utilized to provide high availability between virtual platforms or even between virtual and physical infrastructure.
Here are some benefits of using a unified high availability and backup and recovery solution for all architecture combinations.
- Continuous data replication over any distance and over any hardware, physical or virtual platform provides rapid recovery options for current copy of your data, applications and OS.
- Replicate full server workloads to any data center or disaster recovery location, over standard IP networks, for maximum protection and high availability.
- Implement failover clusters without shared storage or geographic limitations - eliminating the single point of failure and giving you the freedom to locate cluster nodes wherever you want.
- Consolidate workloads with Windows Server 2008 Hyper-V™ or VMware® ESX to reduce costs with a Green IT infrastructure.
- Virtual machine replication from one host to another, in real time and automatically failover those virtual machines in the event of an outage or disaster for high availability architecture.
- Single solution for Windows and Linux availability, supports any application running on those OSs and on either physical or virtual servers.
- Auto discovery protection for business critical systems such as Microsoft Exchange, SQL Server, SharePoint, BlackBerry Enterprise Server (BES), Oracle and more.
All of a sudden this flexibility for a high availability provides IT managers the ability to utilize the infrastructure that is already in the data center versus having to procure new hardware for a disaster recovery solution. This leads to the other trend in 2009 which is recycling for a “Green IT”.
Consolidation and Green IT
Green IT isn’t only about reducing power and consumption and operating more efficiently it is about applying those philosophies to recycling the existing infrastructure in place and making those systems more efficient. One way is to do this is avoid the “rip and replace” solution. If a child has crooked teeth and a dentist presents a solution that requires pulling all the teeth and replacing them with implants you would probably wouldn’t consider that a good option. So, why as an IT manager would you allow a vendor to present a solution that won’t work with the “teeth” you have (a metaphor for existing data center servers) and let them “rip and replace”? I am not suggesting that there isn’t a point where hardware upgrades are necessary to become more efficient but merely suggesting any new infrastructure that is presented should work in combination with the existing high availability solutions that are in place for protecting that same infrastructure.
There are several misconceptions about high availability as well as backup and recovery solutions and much of that has to do with hardware or platform vendors. Just because you need to implement a new SAN or virtual server farm doesn’t mean their solutions is the only one that will work. The truth is it will work but likely for only that architecture and or virtual platform and won’t take into consideration any other data center infrastructure that doesn’t fit that model. A good example of this is today virtual solutions will only work with their virtual platform and can be limited to a single location not providing the ability for offsite disaster recovery. Another example is hardware replications devices that only speak between themselves, which usually requires the purchase of two similar devices so they have a friend at the disaster recovery facility to talk to.
The benefits listed above of a unified high availability solution is innovated to be completely independent of the hardware or virtual platform and provide the ability to backup, recover and provide highly availability solutions across all servers, storage and virtual infrastructure. As data centers continue to consolidate many will be assimilated into a single architecture but there will always be that “red headed step child” of a server or application solution that doesn’t conform but is critical to be shown some love. So before deciding on this is the data protection solution we need, look around and see if that same solution can be used for all the other servers in the data center. Believe me, I have helped many companies recover their data centers and the last thing you want is different procedures for variation of server architecture you might have restore under immense pressure. Keep it simple, keep it unified and keeping it flexible will save money, time and heartache down the road.