Geo-clusters are something that I often get asked about, especially from clients who are looking to protect mission-critical applications and mitigate the chances of data going missing. In this post, we’ll analyse what they are and what they can be used for.
What is a geo-cluster and how can it help prevent data loss?
In order to address what a geo-cluster is, it is first important to understand the concept of a Database Availability Group (DAG). A DAG allows an organisation to have up to 16 replications of an Exchange Database (EDB). Where we can see this come into play is in a situation (e.g. server failures, offline server) where users are prevented from accessing the primary Exchange server. A more detailed explanation of potential scenarios and how to implement DAGs can be found here. Another important term to understand is High Availability (HA), which Microsoft defines as “the implementation of a system design that ensures a high level of operational continuity over a given period of time.”
With both of these terms in mind, we can define a geo-cluster as a type of HA cluster that provides for the physical separation of cluster nodes, meaning that a company with a several locations can physically separate their cluster nodes by placing one primary active server in a primary location (e.g. a company’s headquarters) and replicating the entire Exchange EDB to a secondary passive server in a secondary location (e.g. satellite office, factory). This technology is available with Microsoft Exchange 2007 and is known as Cluster Continuous Replication (CCR). This can be an expensive and complex process to implement, as can be noted from the recommended checklist of requirements that Microsoft provides for it.
Fig. 1. Basic deployment of CCR [Source: Microsoft Exchange TechNet]
This is not available in with Exchange 2010; however, the CCR technology can be seen as Continuous Replication Technology (CRT) in the DAG of Exchange 2010. In this case, only one copy of a mailbox is available at any given time; should this database be rendered unavailable, a DAG component (Active Manager) automatically makes one of the other copies active.
Geo-clustering is considered as an expensive option for organisations to implement in their storage technologies. It is normally used for critical databases, network file sharing and other mission critical applications. It does not appeal to smaller organisations due to the associated costs and larger organisations are usually hesitant to commit to this cost when there are other more secure third-party replication and disaster recovery failover options available to them.
One of the main benefits that can be seen from it is the reduction of risk in the situation where there is damage to the server in the primary location, which can be caused by a variety of things such as fires, floods and other natural disasters. In these cases, the system would be able to rely on its secondary servers, reducing the risk of data loss and downtime.
With the adoption of Office 365, this will also change the way organisations can have guaranteed access to their backups in the event of a disaster recovery. The backup solution of Office 365 is described by Microsoft as:
“Content is replicated from a primary data centre to a secondary data centre, meaning replication is constant and user information will always be backed up in the cloud. Data is stored in a redundant environment with robust backup, restoration, and failover capabilities. This means that – in the event of a technology disaster, theft, fire or flood at the customer’s premises – data, files and documents can be rapidly recovered to ensure business continuity.”
This can solve the issue of implementing geo-clustering, but the main issue here is determining whether this is cost-effective for your organisation as the pricing for hosted solutions is quite different when compared with standard Exchange solutions. However, all the potential costs to your organisation should be considered, e.g. management costs, technical support costs, data recovery costs, backup costs, etc.
It’s all very well having a solid backup plan in place, but what if the worst should happen and data was lost? In my next post, I’ll analyse why an Exchange Server Store size is important in a disaster recovery scenario.
Remember to leave a comment in the section below; I’d love to hear your feedback!