High availability for remote sites
 Remote high availability deployments based on physical or virtual server environments have been made easier with the most recent version of Windows Server 2008 and Hyper-V.
The new remote clustering feature in Windows Server 2008 changes the game. With multi-site clustering, fully remote inter-site clusters can be deployed over the WAN, based on physical or virtual servers. The advantages for high availability and disaster recovery planning are numerous. Previously Microsoft Cluster Services (MSCS) requirement for shared physical disk storage on the LAN limited clustered systems to simple local failover – or at best, a single VPN-based cluster between a pair of sites – a far from adequate solution for robust, large-scale business continuity planning.
Physical and Virtual Infrastructures
Before Windows Server 2008, every server in a cluster had to be on the same subnet. With the networking code completely re-written, servers can now be situated on separate networks. Inter-node communication has been reinforced and is rendered even more robust by ensuring each node is connected to two separate IP networks, for redundancy. According to Microsoft's explanations each server builds its own routing table from the connection information it detects during cluster initialisation. Hyper-V is fully integrated with Windows Server 2008, offering a whole new range of possibilities. Virtual Server R2 offered a somewhat limited range of features. Now, a whole new level of granularity in virtual server clustering is possible, as well as sophisticated load-sharing between virtual servers on multiple hardware platforms within the cluster. In fact, the elemental resource of a a Windows Server 2008 cluster is the virtual server, such that it's possible to attach each virtual server to a specified cluster, so that the physical node on which a specific virtual server runs is predefined. For the enterprise, the result is rapid, cost-effective deployment of large scale disaster recovery/business continuity configurations.
|
|
Don't let data become your weakest link
A key element in any business continuity plan or remote high availability deployment is data storage. Where inter-site clusters are concerned, replication of entire disk storage racks is simply not feasible. Several vendors have looked at how to resolve the storage problem to optimise Microsoft's multi-site clustering offer. SteelEye, for example, offers a solution which eliminates data duplication and interfaces with the multi-cluster services of Windows Server 2008. NeverFail offers data replication for Exchange, SQL, and Sharepoint servers. Double-Take raises the bar with its Geocluster solution, which manages cluster-to-cluster replication in both physical and virtualised cluster environments. Double-Take's Geocluster is the only solution on the market delivering real-time data replication between two nodes using local disk storage. The need for shared storage or costly SAN replication is eliminated, reducing costs and complexity.
Geocluster : A closer look...
|
|