Back in February, Steve Will interviewed Steve Finnes about PowerHA in the blog, PowerHA and Zero-Downtime Backups on IBM i.
That prompted some questions about PowerHA for the smaller shops. I’ve heard many clients state that PowerHA is too complex or expensive for the small-to-medium IBM i environments, so this week, Steve Finnes is here on ‘i Can’ to answer a few more questions.
Dawn May (DM): Why should smaller clients be interested in PowerHA?
Steve Finnes (SF): The top two reasons that our smaller customer implement PowerHA are that it is inexpensive and simple to use.
Oftentimes it is assumed that you need external storage to implement PowerHA, but in fact, it works quite nicely with internal disks and the native host-based geomirroring replication. Customers in this group typically have one or two cores in production and under 1 or 2 Tbytes of storage.
DM: What about cost and complexity?
SF: If you haven’t studied the cost, along with possible savings, it may surprise you. I have business partner friends who tell me that it’s too cheap. For example, let’s take a small shop with one core in production. That means one PowerHA entitlement on the primary and one on the target system. At $3,400 per core, U.S. list price (for async mode), the total PowerHA HA/DR price is $6,800. And note that geomirroring is included in that price. In addition, when you consider that you don’t need a resident expert to manage the legacy replication technology, you’re looking at significant savings. Your only ongoing cost is the usual SWMA subscription, which is 20 percent of the list price for a one-year renewal.
[As for complexity] PowerHA with geomirroring is all about simplicity and certainty of outcome. Once set up, our customers find it easy to use with little day-to-day maintenance, and most importantly, it actually works. One of the comments we hear repeatedly from clients that have moved off of logical replication solutions is that they felt uncertain that everything was there and ready for a failover to the backup system. With PowerHA and geomirroring, these clients will enter into an entirely new paradigm. The replicated data is actually mirrored via the IBM i storage management subsystem, so think of it as being more like disk-level mirroring between two remote locations.
DM: Do we have any client references for these implementations?
SF: We have about 1200 small shops running PowerHA with geomirroring around the world. IBM has published references, and also, given that most of the geomirror installs are done by business partners, those business partners will definitely have references. You should shop around for a business partner that works with PowerHA or give me call for a list of options.
DM: What about skills?
SF: There are two types of skills: the skill required to implement a PowerHA cluster and the skill to use it.
The business partner community with the proper implementation skill set is definitely growing. The knowledge required is a combination of setting up the cluster from the application/system perspective, having the proper configuration, and understanding the network. The best partners out there have these skills and seldom have problems. A business partner that looks at the entire solution end to end is the type of partner that you’ll want to work with. For example, in addition to knowing your bandwidth requirements, you also need to know about the quality of your network. A high bandwidth on a low quality network is a problem. Experienced partners will look at all of the necessary factors.
As for the user skill, once a PowerHA cluster is set up, our customers find it simple to use. There’s the command line option as well the GUI option. The thing about the PowerHA geomirror solution is that there isn’t a lot for the customer to do in the way of ongoing management or day-to-day maintenance.
DM: What kind of general advice do you have for customers looking at PowerHA?
SF: Work with a knowledgeable partner or IBM Lab Services. You want someone who looks at your entire environment. They will study the network requirements, they will look at the primary system and the secondary system (usually a CBU) configurations and they understand how to implement the IASP environment. Partners just learning this can hire Lab Services to do the first few installs. It’s a great way to learn. Also, customers and business partners alike should get a hold of the latest set of redbooks that we just published. Refer to http:///www.redbooks.ibm.com/abstracts/sg247994.html.
Attend COMMON this May where there are several sessions on PowerHA, complete with live demonstrations. It’s a great way to learn and to make contacts with the right people. Refer to www.common.org.
This blog post was originally published on IBMSystemsMag.com and is reproduced here by permission of IBM Systems Media.