Building a Disaster-proof Data Center with HP Serviceguard for Linux, June 2007

Differences Between Extended Distance Cluster and Cluster Extension

The major differences between an Extended Distance Cluster and Cluster Extension are:

• A key difference between extended distance clusters and HP Cluster Extension is the data

replication technology used. The two basic methods available for replicating data between the data

centers for Linux clusters are either host-based or storage array-based. Extended Distance Cluster

always uses host-based replication (MD software mirroring on Linux). Any combination of

Serviceguard supported Fibre Channel storage can be implemented in an Extended Distance

Cluster. Cluster Extension always uses extremely robust array-based replication and mirroring, and

requires storage from the same vendor in both data centers (that is, a pair of HP StorageWorks XP

disk arrays with Continuous Access, or a pair of HP StorageWorks EVA disk arrays with

Continuous Access).

• Data centers in an Extended Distance Cluster can span up to 100 km, whereas the distance

limitations between data centers in a cluster with Cluster Extension are defined based on the

storage array.

Cluster Extension EVA distance supported is defined by the shortest of the following distances:

– Maximum distance that guarantees a network latency of less than 20 minutes

– Maximum distance of 500 km distance

– Maximum distance supported by the data replication link

– Maximum supported distance for DWDM as stated by the provider

Cluster Extension XP distance supported is defined by the shortest of the following distances (In

contrast to Cluster Extension EVA, there is no Cluster Extension software limiting factor.):

– Maximum distance supported by the data replication link

– Maximum supported distance for DWDM as stated by the provider

• In an Extended Distance Cluster, there is no built-in mechanism for determining the state of the data

being replicated. When an application fails over from one data center to another, the package is

allowed to start up if the volume groups can be activated. With a Cluster Extension implementation,

an application is only allowed to start up based on the state of the data and the disk arrays.

Data might be updated on the disk system local to a server running a package without remote data

being updated. This happens if the data link between sites is lost, usually as a precursor to a site

going down. If this occurs and the site with the latest data then goes down, that data is lost. This

time from losing the link to the site going down is known as the recovery point. An objective can be

set for the recovery point so that if data is updated for a period less than the objective, automated

failover occurs and a package starts. If the time is longer than the objective, then the package does

not start. In a Linux environment, this is a user configurable parameter: RPO_TARGET.

• Extended Distance Cluster disk reads can outperform Cluster Extension in normal operations.

However, Cluster Extension data resynchronization and recovery performance are better than

Extended Distance Cluster.