Understanding and Designing Serviceguard Disaster Recovery Architectures

4 Comparison of Disaster Recovery Solutions
Table 7 summarizes and compares the disaster recovery solutions that are currently available:
Table 7 Comparison of disaster recovery Cluster Solutions
ContinentalclustersMetroclusterExtended Distance Cluster
for RAC
Extended Distance
Cluster
Attributes
Increased data
protection by
Two significant benefits:Excellent in “normal”
operations, and partial
Excellent in “normal”
operations, and
Key Benefit
• Provides maximum
data protection. State
failure. Active/active
configuration provides
partial failure. Since
all hosts have access
supporting unlimited
distance between data
of the data is
maximum data throughto both disks, in a centers (protects against
determined before
put and reduces the needfailure where the node such disasters as those
application is started.
for failover (since bothis running and the caused by earthquakes
If necessary, data
data centers are active,application is up, but or violent attacks,
resynchronization is
the application is already
up on the second site).
the disk becomes
unavailable, no
where an entire area
can be disrupted).
performed before
application is brought
up.
failover occurs. The
node will access the
• Better performance
than Extended
remote disk to
continue processing.
Distance Cluster for
resync, as replication
is done by storage
subsystem (no impact
to host).
No automatic failover
between clusters.
Specialized storage
required. Currently, XP or
SLVM configuration is
limited to two nodes for
No ability to check the
state of the data
Key Limitation
P9000 with Continuousdistances of up tobefore starting up the
Access, EVA or P6000100km*. CVM or CFS,application. If the
with continuous access,which are available withvolume group (vg) can
HP 3PAR Remote CopyServiceguard Storagebe activated, the
and EMC’s Symmetrix
with SRDF are supported.
Management Suite
Bundles, configuration
application will be
started. If mirrors are
supports up to eightsplit or PV links are
nodes. However,down, as long as the
eight-node configurationvg can be activated,
is limited to a distance ofthe application will be
10km*. Datastarted. Data
synchronization canresynchronization can
create an impact onhave a big impact on
system performance assystem performance,
this is a host-based
solution.
as this is a host-based
solution.
No distance restrictions.Shortest of the distances
(up to 300km) between:
* 100km (maximum is
two nodes, with either
CFS, SLVM or CVM)
* 100 KilometersMaximum
Distance
• Cluster network latency
(not to exceed 200ms).
* 10km (maximum is two
nodes with SLVM and
• Data Replication Max
Distance.
eight nodes with CVM
and CFS)
• WDM provider max
distance.
You have a choice of
either selecting your
Array-based, via
Continuous Access XP or
Host-based, via
MirrorDisk/UX or (Veritas)
CVM and CFS
Host-based, via
MirrorDisk/UX or
(Veritas) VxVM.
Data
Replication
mechanism ownContinuous Access EVA
Replication can impact
performance (writes are
Replication can affect
performance (writes
are synchronous).
Serviceguard-supported
storage and data
replication mechanism,
or 3PAR Remote Copy or
EMC SRDF.
Replication and
resynchronization
synchronous). Re-syncs
can impact performance
Re-syncs can impact
performance (full
or implementing one of
HP’s pre-integrated
performed by the storage(full re-sync is required in
70 Comparison of Disaster Recovery Solutions