Some discussion of Andy Tai's 'The Software Wars'

This is about The Software Wars, a graphical look at the emergence of Open Source software, especially the latest, from 2000 thru 2006. Wikipedia and others agree with your instructor's opinion that these graphics provide an excellent overview of the situation of IT legacy and emerging IT. There's hardware involved too, and manufacturers, but hardware is less important these days, except for those locked in to a proprietary platform, it is bargain & commodity priced.

Some stuff is missing: The earlier diagrams only draw in Desktop and Server-class battles where Microsoft was fighting for market share, and express some optimism that Linux might make significant victory in the desktop market. The later diagrams add in some hint of larger-class hardware, too, naming or hinting at a couple instances of the mid-range and mainframe classes that are so important for much enterprise today, where Microsoft is continuing a successful fight for market share. And, a lot has happened since 2006 that needs comment while we're waiting for another graphical update. The newest graphic, 2011, is not really a continuation of the Andy Tai series, but it does provide a good look at what's going on in the Personal Computing marketplace.

Two Major Markets in the Battle Zone

What's important for a student of IS to know is there are two major 'markets' in business computer technology, today expressed as: desktop and server. Everybody's very familar with the desktop/notebook/tablet/pad that provides an excellent Personal interface for employees at a desk or in the field. Any of these can reach out and grasp the corporate data stream, read from it, inject into it. On business desktops, Windows has 90% of the machines firmly in its grasp. Only network admins and the CIO are very familiar with the server/mid-range/mainframe machines that run the business. These machines have traditionally been very _not_ Windoze!

The back-office/website of any long-lived enterprise, department, or agency that computerized in or since the 1960's runs on mid-range or mainframe hardware. Windows is new in this environment! Only recently, as at 2010, has Windows reached anywhere near %50 of the 'server market'. There, 'Unix machines' have been rock-solid since the '70s, continuously gaining market share, continuously improved, always state of the art in performance & reliability, and always so very _not_ Windoze! Google does _not_ run on Windoze! Neither does 'the government', although most government employees' desktops do.

Combatants at the Battle Lines

The major combatants implied in the Server Operating Systems battle lines drawn in Mr Tai's diagram are application software developers, network admins, and users from these camps:

What we're seeing in the Software Wars is mostly about how Windows is competing against these other three camps. With a maniacal marketeer at the center, assimilation is only a matter of time, isn't it? But, he retired...

Windows _is_ gaining share in the large-system market for good reasons and is a good OS. We'll be seeing it get more share in network rooms and back-offices as Micro$oft wraps itself around the huge RAMs that can be utilized with 64-bit CPUs. MS was still learning lessons that UX has mastered for decades, but with all the open source solutions for Huge RAM and Huge Disk Farms there for anybody to see, we expect MS will master this stuff very quickly. MS's climb to about 50% share in the server market isn't showing that 'hockey stick' acceleration yet, might never...

The Rub so Far: Mid-range and Mainframe machines' 64-bit CPUs have for decades allowed them to reference TBytes of RAM and provide millions of concurrent users with sub-second response time from one machine. Windows servers have only recently been deployed on 64-bit Intel/AMD so all prior, 32-bit, versions have been limited to a max 4 GBytes of RAM per server, so any 'scalability' has come by deploying servers in large farms, or blades, with each server supporting perhaps a few hundred to a thousands of concurrent users. 'WinTel' 64-bit workstation/server mainboards can, at 2011, accomodate RAM of 32 GBytes thru 128 GBytes and service proportionately more users. In comparison, an IBM zSeries can accomodate a couple dozen 'CPU Books' with Seven CPUs each, each Book having 792GBytes of 'local RAM' + there can be multi-TBytes of RAM deployed in mirrored-arrays available to any of the CPUs in the Books. RAM operates hundreds-of-thousands to a million times faster than Hard Disk, so mainframes and mid-range machines have in the past had this competitive edge in the large-sytem market.

Rule of Thumb at Windows/Legacy scirmishes ca. 2011: one big mid-range computer can handle the load of a thousand or more Windows servers; one big mainframe can handle the load of a few thousand Windows servers.

Not Shown Clearly in the Diagrams, but In the Vast Legacy

Really 'Big Iron', mainframes from IBM, Sun, and a few others, are missing entirely from the earlier diagrams in The Software Wars, are hinted at on the last, at 2006, and they are very important in 2011 since these big machines have major market share in enterprise-class and government computing. Also 'IMS' needs to be drawn into the diagram, which is IBM's hierarchical DBMS that runs on their mainframes. And Oracle, an Enterprise-class RDBMS that often runs on Sun mainframes. Together, IMS & Oracle manage something more than 90% of the worlds' data, so they are very significant in the overall IT marketplace but are sometimes overlooked when the world's data are contemplated by noobs. UniVac mainframes and HP mid-range machines and their CODASYL DBMS survive today as another important environment used extensively in government systems, airlines, cattle ranching, and other large enterprises that were mainframe-based back in the '70s when UniVac was one of IBM's competitors.

Perhaps these can be thought of as 'those lands down there' that have until recently been entirely out of the warzone centered by the dark Emperor Bill as he reconnoiters the hills of the fervent UX and mainframe people who have for decades cared for the data of enterprise? All this is up for grabs in the next decade, as a generation who has grown up with Windows starts making decisions about what will be in the IT legacy in 2050...

Here's roughly what was covered in lecture about these diagrams: