Readiness Assessment Accelerates Moving IBM i Workloads To The Cloud
November 7, 2022 Timothy Prickett Morgan
There are on the order of 120,000 unique IBM i customers in the world, and the one thing they all have in common is that they are absolutely – each and every one of them – unique. There are no two customers alike because they have lots of homegrown code, or a mix of their own code and that of an ISV, or maybe they have a heavily customized stack from an ISV.
We defy you to find one customer – even one – that is running any software from a third party as it is, out of the shrink wrap and into the Power Systems box, unchanged. There are probably not many customers with similar, much less the same, configurations of servers, storage, networking, and tooling to create and manage those applications. Even if they are the same size company, even if they are in the same industry, and even if they are running the same third-party software.
AS/400 and IBM i professionals believe – and correctly so – that the value they bring to their companies is the customization of hardware and application software, fit for purpose and built for the long haul. That extreme diversity in the IBM i base, however, presents some problems for customers who want to move to the cloud as well as the cloud service providers who want to move them from on-premises systems to elastic capacity with utility pricing off premises.
“I sold Power Systems hardware for a long time,” says Jeff Swartz, vice president of sales for the Fresche Cloud business unit of Fresche Solutions (Abacus Solutions). “Cloud is very different. For most customers in a Power-based cloud, they are going to buy capacity on a shared environment and the beauty of that sharing is that customers can change the compute capacity they use over time, both up and down, and only pay for what they use. You no longer have to pay for what is sitting on the floor, for end of week, end of month, end of year, or other kinds of peaks. So, you can’t just take what you have in terms of core count, CPW performance, gigabytes of memory, and I/O bandwidth and just move it over to a cloud instance.”
Well, you can, but this still doesn’t give you an optimized cloud environment based on your actual needs. Which is the point Swartz is making. “We want our customers to take advantage of the real computing power of cloud – and that starts with an understanding of which applications and workloads make the most sense to move, what capacity is really required and a plan for how that capacity can be spread out over time. This is what the Cloud Readiness Assessment from Fresche Cloud is all about – working with organizations to simplify and accelerate cloud readiness and adoption.
The Cloud Readiness Assessment has a few main goals:
- Define the business goals and objectives for cloud and the desired outcomes
- Assess and analyze the existing workloads for cloud suitability
- Determine which applications/data/workloads will move
- Evaluate infrastructure and security requirements
- Establish the ideal cloud operating model
- Build a cloud budget with future scaling
Swartz adds: “As part of the exercise, we get prospective cloud customers to rethink what they are doing and try to get faster time to answer for as many of their interactive and batch workloads as makes economic sense. Then, we create a cloud infrastructure that meets the business needs and only activates capacity when and where it is needed across the course of a year once this rethinking is done.”
Here is a good example that we talked over with Swartz and Thomas Harris, chief operating officer and general manager of the Fresche Cloud business unit.
Let’s say you have a production IBM i machine in your datacenter with six cores running your applications and your databases. To do your nightly batch workloads – creating and sending out invoices, printing out statements, generating reports for management, creating bills of material for manufacturing, whatever – takes ten hours on this hypothetical machine. During the day, it really only takes two Power9 cores to do the OLTP work and the six cores is really just there for that batch work. Now, move to the cloud. You create a production partition to do that OLTP work that has just two Power9 cores, and everything is good. You add a layer of high availability clustering to the machine, which you have always wanted to do, and pay a monthly subscription for MIMIX or Maxava or whatever to do the replication out to a skinning partition that is running on a beefier machine – say a Power E980 with 1921 cores – but you are only replicating out to one-tenth of a core on this machine most of the time. Right up to the moment when you want to do the invoices and statements, or some complex queries. Then, you massively expand that logical partition on the Power E980 with a replicated copy of your databases, perhaps with several dozen cores underneath it, and all of that batch work completes in an hour. After an hour, you turn that partition back down.
This is the kind of flexibility and agility that the cloud offers, and that is why you need to do more than a simple back-of-the envelope list of basic feeds and speeds on your current machine and fire up the same capacity in the cloud.
“The real concern at IT shops is all around speed to value and time to answer,” explains Harris. “And if you can shrink that batch processing or complex query processing down to minutes or seconds, that’s what the business managers care about. Operationally speaking, such a speedup is great, but what is really great is how such expandable cloud capacity translates to business value. Managers get to make better decisions because they can ask more questions and they can get answers back fast.”
Such system sizing exercises are not new to IBM i shops, of course. Big Blue and its reseller partners who work directly with IBM i shops have been doing this for decades. But this has some twists to it because dynamic cloud capacity is very different from a static on-premises machine.
“We want to be able to do the exact same thing, but the difference is that our assessment is not just for a new on-premise setup for a customer,” says Swartz. “In fact, we are doing assessments for customers who want to move from some of the other IBM i clouds out there. No matter where they are coming from, we want to actually analyze what it is they’re doing today, and there’s some thought that goes into this as opposed to just counting cores, CPWs, and disks and cloning that. This Cloud Readiness Assessment is done with a combination of our pre-sales engineers and cloud architects as well as some automating tooling that we have created that puts together a detailed solution for the customer and a proposal from Fresche Cloud.”
Fresche Cloud is charging a nominal fee for the Cloud Readiness Assessment, but it is a nominal charge compared to the ongoing costs of buying and running an IBM i system or buying capacity on a cloud for one, two, three, or five years. This assessment, by the way, is not designed to be a budgeting tool or a return on investment tool. There are far too many nuances at IBM i shops to easily generalize and categorize costs. But, the output of this assessment certainly could be a key part of an ROI analysis.
The following is perhaps the most important reason why customers should do a Cloud Readiness Assessment – even if they decide not to move to the cloud right now.
They are probably going to move at some point, and there are a lot of reasons for that, not the least of which is that companies do not want to invest in hardware and the people to manage these systems, particularly when every passing year reduces the pool of IBM i experts in the world as they retire. Even when customers get the budget together to buy new Power Systems so they can expand their IT operations to help move the business forwards, they will still be struggling to maintain and manage and secure those systems when what they really need to be doing is working with organizations like Fresche who have the skills and ability to modernize applications and creating new ones. Sooner or later, push will come to shove here, and doing an assessment now might give food for cloud thought as your company is wrestling with such challenges.
Friends don’t let friends buy hardware or manage hardware they don’t need anymore. They let the clouds handle that.
Fresche currently supports hundreds of customers and manages thousands of IBM LPAR’s of all sizes. They also have a unique ability to support large enterprise P20 and P30 servers. Organizations interested in the assessment or in talking to Fresche about IBM cloud can reach out to Jeff Swartz at email@example.com or firstname.lastname@example.org.