If we set our sights on achieving the most cost-efficient solution, then it may not be true that the solution with the lowest initial purchase price wins.
For example, take flash storage. For years it has been easy to show that when you include in your evaluations various operational savings, productivity gains, reliability comparisons, maintenance expenses, and the list goes on – flash offers lower overall costs than mechanical hard disks. In fact, it’s not even close. Yet, still today you can find enterprise-level data storage users and decision-makers talking about the higher cost of flash – because the initial purchase price of flash storage is, indeed, higher than the lowest-priced hard drives.
Total Cost of Ownership (TCO) is the industry term for considering all the costs of a solution. Here is how TCO is described by a popular IT information source.
“Total cost of ownership (TCO) is an estimation of the expenses associated with purchasing, deploying, using and retiring a product or piece of equipment. TCO quantifies the cost of the purchase across the product’s entire lifecycle. Therefore, it offers a more accurate basis for determining the value – cost vs. ROI – of an investment than the purchase price alone. The overall TCO includes direct and indirect expenses, as well as some intangible ones that may be assigned a monetary value. For example, a server’s TCO might include an expensive purchase price, a good deal on ongoing support, and low system management time because of its user-friendly interface.”
Recently, the IT industry analysts at Edison Group conducted a TCO comparison between IBM Spectrum LSF – a suite of enterprise and high performance computing (HPC) application workload management solutions – and comparable open source software. The same online information outlet whose definition of TCO we used above also provides us with some thoughts about “open source”:
“In general, open source refers to any program whose source code is made available for use or modification as users or other developers see fit. Open source software is usually developed as a public collaboration and made freely available.”2
IBM participates in many different open source initiatives – for example, providing major contributions to open source Spark – data analytics software that’s soaring in popularity right now.
The Edison TCO comparison between IBM Spectrum LSF and open source workload management software makes a great deal of sense when you note, as Edison did, that “…there is a persistent market perception that open source is always a less expensive alternative to any solution that contains proprietary software.”3 And of course, Spectrum LSF is a proprietary software product offered by IBM.
This particular TCO comparison makes good sense for another reason – workload management is complicated. Early in the TCO comparison, the Edison authors note: “To complicate matters, IT infrastructure is in transition as companies move towards hybrid processing models which include traditional bare metal, virtual machines, cloud and containers…This complex infrastructure cannot function efficiently using simplistic workload scheduling tools. Each type of processing model requires a slightly different approach to achieve the highest level of efficiency, therefore creating potentially conflicting operational focus which a simple workload scheduling tool cannot satisfy.”
Complexity within the environment where a TCO analysis is being conducted means that there will be a number of cost variables involved – a lot more than just the initial purchase price. These are the environments where the concept of total cost of ownership is especially valuable to those who will actually be making purchase and deployment decisions. And HPC workload scheduling and management certainly qualifies in this regard, requiring job prioritization and scheduling, operations management, workload and resource balancing, and even handling of requests for additional resources from the cloud, which then adds complications caused by location of data for example.
Edison notes that the “IBM Spectrum LSF family addresses all of the challenges listed above. It is much more than a workload scheduler. It is a complete solution for HPC resource and workload management, designed for enterprise-grade productivity.”
Edison goes on to list some of the ways that IBM Spectrum LSF provides TCO advantages. The solution:
- Provides a simple and effective User Interface
- Is easy to integrate with all the other tools in an HPC processing environment
- Provides valuable operational analytics
- Automates job workflows
- And even offers intelligent flexing into cloud resources.
These are just a few, but a key area where IBM Spectrum LSF really helps lower TCO is performance. When workloads are scheduled more efficiently, less infrastructure such as hardware is needed. Hardware can be the single biggest cost factor, and as the Edison analysis illustrates, IBM Spectrum LSF can require less of it to achieve similar productivity to that of open source software.
Productivity itself is not a simple concept, potentially spawning multiple cost variables within a TCO calculation. The Edison TCO analysis highlights an interesting cost factor called End-user Productivity Opportunity Cost, and it results in the biggest TCO advantage provided by IBM Spectrum LSF.
What factors did Edison consider when calculating the cost of End-user Productivity Opportunity? You should read the report yourself to get the full explanation. You’ll also want to learn more about all seven of the major cost factors considered in this particular TCO comparison.
And what was the bottom line? Well, we wouldn’t want to give away the very best part of this TCO story. Let’s just say that when it comes to cost, performance, and productivity, IBM Spectrum LSF has a lot to offer.
Read the full report here.
[3] Edison Group: Comparison of Open Source software vs. IBM Spectrum LSF Suite for Enterprise, June 2018 (URL TBD)