Admin Alert: Risk and the Power i Hardware Upgrade

November 3, 2010 Joe Hertvik

While upgrading Power i hardware is a hassle, the benefits in alleviating processing bottlenecks, providing new capabilities, and solving performance issues far outweigh the risks. Before an upgrade, however, management will want to consider any possible upgrade risks and how they can reduce your organization’s exposure to those risks. This issue, I’ll examine some common i/OS hardware upgrade risks and what you can do about them.

Getting the Vocabulary Right

In discussing risks, organizational requirements may require you to specify the following items when requesting monies for a new Power i hardware upgrade.

Can you identify your upgrade risks and how those risks can impact the organization? You may be asked to quantify negative outcomes in dollars, time, or some other tangible measurement. Besides specific details, your organization may also want you to specify what the business impact of each realized risk would be on a simple scale (low, medium, high).
What is the probability that the risk will be realized (i.e., how much should we worry about this happening)? Risk probability can usually be classified as having a low, medium, or high chance of the situation occurring.
If the risk is realized, how will the organization respond to the situation?

Generally, organizations respond to risk in the following ways.

Avoidance–Performing an action in advance to help insure the risk will not be realized.
Mitigation–Lessening the impact if the risk does occur.
Planned Contingency–If the risk can’t be avoided or mitigated, what will the strategy be for dealing with the issue? Think Plan B.

Given these parameters, here are some of the risks, probabilities, and responses I recently compiled for an upcoming Power i upgrade for one of my clients. Planning for these situations will give you more control over your upgrade.

Risk
Description

Potential
Impact if Risk is Realized

Probability
of Occurrence

Response

Downtime risk: Existing
production is off-line for 12- to 24-hours while system is being upgraded or
migrated to new hardware.

High: In 365x24x7 environment,
system will not be available for production, entering orders, or for outside
users. Depending on how long it takes to complete an upgrade, this could be a
significant amount of time.

High: 100% certain if not
avoided or mitigated

Avoidance: If system is
configured with a Capacity BackUp (CBU) system, switch
over to the CBU while upgrading the production box. Note that you may still
experience some downtime while your production machine is being switched over
to the CBU and vice versa.

Mitigation: Notify
customers in advance that machine will be down. If accessed by Web sites,
post warning notices on each site warning of upcoming outage. Perform upgrade
in extremely low-traffic hours, such as Sunday morning or on a holiday.

CBU Processing Risk: When
switching over to the CBU during an upgrade, CBU is not configured correctly
and data is corrupted.

High: If information is not
being replicated correctly, processed orders could contain errors.

If the CBU is not correctly certified against source machine, critical
files or libraries could be missing or contain old data.

It’s also worth noting that many companies have never used their CBU
for live processing or in some cases, have never truly tested it. If that’s
the case, it’s imperative that you certify your CBU for live processing
before switching over.

Low to Medium: Probability
depends on how well tested the CBU is.

Avoidance: Certify the CBU
switchover process to provide confidence that it is properly configured to stand-in
for production during the upgrade. For more information on CBU certification,
see the Related Stories section for my series of articles on certifying a CBU
for live switch over.

Mitigation: Keep
programming and database support personnel on hand during upgrade switchover
to monitor system and correct problems.

Hardware failure risk: New
hardware fails after installation and migration.

High: Production systems
down.

Low: Power i systems are
famously reliable but it is possible there could be a defective component on
the new box or that something could have been damaged during shipping or
installation.

Avoidance: If possible, power
up and run the new hardware for several days to a week before migrating your existing
system over. This will give any defective components a chance to fail before you
start relying on them for production. Pre-migration activation can be
particularly helpful to just let your disk drives spin for a few days to
break them in. Note: you may not be able to do for a system upgrade.

Planned Contingency: Switch
over to CBU, if available.

Financial risk (Leasing): If
leasing, your monthly payments go up on the new machine, perhaps by a
significant amount.

Medium: If you are not
familiar with lease procedures, it may be difficult to determine if the
leasing company is taking advantage of you. Your company may be paying more
than it needs to for the new machine and you may never realize it.

Low, Medium, or High: Controlled
by a number of factors, including the terms of your current lease, whether
you are turning your lease in early, how much equipment is being carried over
to the new lease, etc.

Avoidance: If you don’t
have a leasing background, look to add someone with leasing expertise to your
team, so that you have the firepower to examine and negotiate hard with your
leasing company. A good leasing negotiator may be able to help your
organization save tens of thousands of dollars (or more) on a lease.

Avoidance: Don’t place
your lease with your business partner. This can create a conflict of interest
that could lead to increased leasing costs.

Avoidance: Get competitive
quotes on your lease, no matter how long you’ve dealt with your current
leasing company (especially if you’ve dealt with the same leasing company for
a long time).

Avoidance: If you are
upgrading your current machine and using existing components to cut the
equipment cost for the new configuration, find out what percentage of the new
leasing cost is devoted to the Base Lease Extension (BLEX) of the reused
equipment. The BLEX is the value of the equipment from the old lease that is
not being returned after the lease ends. For reused equipment, the residual
equipment value is rolled into the new lease. The BLEX is important because
it can inflate your new lease cost and it can sometimes result in you paying
a second time for equipment the leasing company has already paid for (pure
profit for the leaser). You may be able to negotiate and lower the BLEX
component of your new lease, because the BLEX is merely the value that the
leasing vendor has assigned to the reused equipment. That value may or may
not be fair and accurate, and it is entirely set by the leasing company.

Avoidance: Have the
leasing companies break down your quote into individual component detail, so
that you know which lease items are more expensive. This will enable you to
see where the vendor is making his money.

For a good starting point on Power i leasing issues, check out my
earlier article getting the best deal on a lease for an i/OS machine in the
Related Stories section.

Operating System Compatibility
risk: You are currently running i5/OS V5R4M0 but are looking at upgrading
to a Power 7 machine. V5R4M0 does not run on Power 7

Medium: If you’re not
planning on an i 6.1 or i 7.1 upgrade before installing your new hardware,
your upgrade options will be limited to Power 6, which can run V5R4M4.

If you upgrade your partitions to 6.1 or 7.1 before installing new
hardware, you will have to deal with all the risks of a new upgrade,
including the possibility that critical software won’t run on the new OS,
increased expenses for software licensing, and possible consulting fees to
help with the upgrade.

Also note that if you upgrade, you will face a downtime risk where
the machine will be unavailable for a set period of time or you will need to
switchover to the CBU while the upgrade is occurring.

High: You must perform an
upgrade if you want to migrate to Power 7 hardware

Low: If you want to limit
your hardware to Power 6 machines, you can continue to run V5R4M5.

Avoidance: Don’t upgrade
your operating system at this time and migrate to a Power 6 machine. This
would be dependent on finding a Power 6 machine that can be sized correctly
for your workload and that can grow with your workload over the course of the
lease.

Mitigation: Upgrade your
partitions to i 6.1 or i 7.1 so that you can purchase Power 7 hardware and
migrate your existing systems to the new hardware.

Mitigation: If upgrading
your operating system, consider going to I 6.1 instead of 7.1, because 6.1 is
more established, more third-party vendors will probably support it, and i 7.1
is relatively new, which can lead to problems.

Keep in mind that no upgrade is without risk. However, once you know and identify the risks, you’ll have more knowledge about how to plan and deal with any risky situations you’ll face during the upgrade cycle.