Ongoing Saugatuck dialogue with enterprise IT and business leaders, and with IT providers, suggests that we are on the cusp of a significant shift in business application usages and a resulting shift in how servers (and related offerings) are developed, delivered, and utilized within enterprises and within Cloud providers’ platforms.
It’s not completely new in IT (what is?), but it does appear to indicate a rapidly-accelerating trend that will increasingly shape and re-shape enterprise IT buying patterns, and IT (including Cloud) providers’ business strategies and offerings, for years to come: Servers and software fully capable of general purpose workloads, but also including accelerators and optimizations for specific current and potential/likely future workload types.
Saugatuck projects that savvy vendors, in pursuit of market share, will introduce and heavily promote an increasing number and range of such workload-optimized offerings for enterprise in-house infrastructures and also for Cloud-based solutions.
Why is it Happening?
Optimizing performance based on specific workload requirements is certainly not a new phenomenon. Beginning in the mid-1970s mainframe vendors recognized the competitive advantage of optimizing the design of their offerings based on the frequency of usage of instructions by specific workloads (e.g., scientific, commercial batch, IMS, CICS, etc.). Not surprisingly, over the ensuing years, designing for typical workloads became “the norm” for vendors of servers, storage, networking offerings, operating systems, and middleware.
For the past several years, fundamental characteristics of workloads, and how they are managed, have remained relatively stagnant. Meanwhile, similar to the focus in mainframes in the early days of virtual machines, there has been intense design focus for x86 servers in optimizing for virtualization: some instructions have been accelerated, other instructions and hardware functions have been created, etc.
However, based on our past several global enterprise research surveys, as well as our work with leading IT providers, Saugatuck believes that the Cloud-catalyzed, expanding adoption of Analytics, Social IT, and Mobility is causing rapid acceleration in both the evolution and the scope of enterprise IT workloads. Traditional workloads are growing in number and scale, while new types of workloads (often engendered by rapidly-expanding use of Analytics, Social IT and Mobility) are emerging and growing even more rapidly.
The phenomenon is explainable by the Jevons Paradox: When individual items, services, pieces or actions become less costly, our volume of consumption of those increases, often becoming part of a more active and complex environment. Mobile phones, social networks, and cable TV are good examples; consumers spend thousands of dollars individually per year on each, because of the perceived low cost of acquisition and of discrete uses. And the more we use these items or services, the more we become reliant upon them and the more discrete uses (i.e., workloads) we develop or invent, and so the more complex and costly the entire environment becomes.
This often-accelerating cycle of growth and complexity both enables and drives increasing demand for solutions that are both cost-effective and efficient increasingly, this means solutions that are optimized for specific workloads, or for workload types/categories.
Click here to read the market impact.
This blog was originally published at Saugatuck's Lens360 blog on January 9, 2014. Published with permission.