Getting started with the IBM® DB2® pureScale® Feature is
quick and simple: Applications do not have to be aware of the topology
of your database environment when you deploy the feature. This means
that applications work just as they did before, yet they can benefit
from the extreme capacity and continuous availability from
the moment that you start your DB2 pureScale instance
for the first time.
Increasing capacity
Capacity planning with
the DB2 pureScale Feature is
simple. You can start small and add members to
your database environment as your needs grow, scaling out from the
most basic highly available configuration all the way to the maximum
supported configuration, which provides extreme processing capacity.
Scaling is near linear in efficiency and highly predictable.
When
you scale out, no application changes or repartitioning of your data
is required. No performance tuning is required to scale efficiently.
If you need more capacity, you simply add more members.
Maintaining availability
Maintaining database
availability means both compliance with service level agreements (SLAs)
and high tolerance to component failures. To maximize hardware utilization
rates and help keep response times consistent for your applications,
incoming database requests are automatically load balanced across
all active members in
your DB2 pureScale instance.
To minimize the impact of component failures, the automated restart
and recovery process of the DB2 pureScale Feature runs
quickly and without affecting most database requests. Only those database
requests that were being processed by a failed member must
be resubmitted by the originating application; the resubmitted requests
are then processed by the next available member.
In
the example in the following diagram, several events take place in
short succession. Multiple component failures require automated, internal
recovery, and a scale-out operation increases the capacity of the
DB2 pureScale instance.
None of these events requires any application awareness. The
box containing the components of the
DB2 pureScale Feature is
shaded to indicate application transparency.
Figure 1. A DB2 pureScale environment
encountering multiple component failures and being scaled out. Applications
connecting to the database need not be aware of these events.
Planning made simple
The ability to easily
add and remove resources helps you to manage challenges such as the
following ones:
- Cyclical workloads. If some of your workloads are cyclical (for
example, seasonal), you can add resources before they are required,
and then move the extra capacity somewhere else later on.
- Sudden increases in workloads. An SLA might dictate minimum response
times for completing database requests. If you discover sudden workload
surges from some applications that are threatening response times,
you can help meet your SLA by quickly moving additional members to
the database that is experiencing the peak demand.
- Maintenance-related slowdowns. To help negate the effect of system
maintenance on the overall throughput of your DB2 pureScale environment,
you can add amember to
your environment before commencing maintenance on an existing member.
After you complete the system maintenance and the original member rejoins
the instance, you can remove the additional resource or perform maintenance
on other members.