Feature Story


More feature stories by year:

2024
2023
2022
2021
2020
2019
2018
2017
2016
2015
2014
2013
2012
2011
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998

Return to: 2014 Feature Stories

CLIENT: SUVOLA

July 15, 2014: Computing Now

Computing Appliances to Accelerate Deployment of New Hardware and Software Technologies

By Christopher M. Piedmonte and David L.R. Stein

The advent of hyperscale semiconductor integration (HSI) is now driving development of a new generation of massively parallel processing (MPP) computer systems for enterprise computing.

These so-called "microserver" systems are needed to bring the many combined benefits of HSI and MPP to enterprise computing. Unfortunately, to date HSI has been limited largely to reduced-instruction-set computer (RISC) architectures for which there is little software infrastructure applicable to enterprise computing.

Thus, there is a significant gap between the currently available software infrastructures for these RISC architectures and that which is required for general-purpose enterprise computing. Until the gap narrows, the many combined benefits of HSI and MPP will be denied to enterprise computing, except via new special-purpose peripheral-computing appliances–comprising all the application-specific hardware and software required to perform legacy enterprise-computing functions–that are specifically designed to transparently interoperate with legacy enterprise-computing systems via de-facto standard interfaces.

Until recently, HSI has been driven by mass-market demand for lower-power, lower-cost, and higher-performance consumer electronic devices like electronic games and smartphones. Simultaneously, MPP has been driven by continuing demand for increased computing capacity and the inability of semiconductor technology to produce ever-more-powerful uniprocessors with acceptable thermal-design power characteristics.

With the singular exception of big data applications, enterprise computing has made limited use of multiprocessing to date–to say nothing of MPP, which previously has been exploited primarily for purely-scientific or high-performance computing. While a new software ecology is currently evolving around big data applications, that ecology is largely based on open source software incapable of running legacy enterprise-computing applications.

Currently, only "innovators" and "early adopters" can cope with the complexities and vagaries of multivendor installations of constantly evolving, inherently unstable and nonstandard open source software, the currently evolving big data software ecology addresses only a small part of the enterprise IT infrastructure required by a new generation of massively parallel, general-purpose enterprise-computing systems.

Consequently, the trends to HSI and MPP for general-purpose enterprise computing bring with them a requirement for a new enterprise IT infrastructure. This new infrastructure can be broken down into two symbiotic parts–a new software ecology to transparently run legacy enterprise-computing applications in the new environment, and a new hardware and software ecology to transparently bring the many combined benefits of HSI and MPP to peripheral subsystems.

Creating a new software ecology is more difficult. Until the gap between the software infrastructures for the most popular legacy complex-instruction-set-computer architectures, and those for contemporary RISC architectures can be closed, the new microserver systems will be unable to run the vast majority of legacy enterprise-computing applications.

Fortunately, peripheral subsystems have no requirement to run legacy enterprise-computing applications, per se, but require only the capability to interoperate with them via de-facto standard interfaces. Thus, new special-purpose microserver appliances comprising all the application-specific hardware and software required to transparently interoperate with legacy enterprise-computing applications will be deployed years before availability of general-purpose microserver systems capable of running legacy enterprise-computing applications.

Christopher M. Piedmonte is CEO and cofounder, and David L.R. Stein is president and cofounder, of Austin-based Suvola Corp. The company provides enterprise infrastructure software specifically designed for the SoC-based cloud infrastructure market in the areas of systems management, secured Web application delivery services and big data processing for strategic and tactical business analytics. Chris is a serial entrepreneur with a 30-year track record of successfully creating innovative state-of-the-art solutions to a wide variety of problems in complex information technology systems. Dave is a computer industry and startup veteran who also cofounded global market research firm Gartner as well as a $250 million venture-capital partnership. Contact Chris at chris@suvola.com, and Dave at dave@suvola.com.

Return to: 2014 Feature Stories