What is Big Software?
At Bloor Research those of us in the infrastructure practice are getting evangelical about Big Software. However, the term prompted a bemused question, that is the title of this article, from our esteemed Research Director for Information Management, Philip Howard. The flurry of quick and short email answers from colleagues did little to ease Philip's bemusement and, dare I say it, his humour.
As an industry we are often fast and loose with our naming conventions and descriptions, often allowing them to become misused and corrupted over time. Why "data networking" and not just "networking"? Why are network applications, qualitatively, any different from an HR application? Software-defined storage being corrupted to storage as software; does that mean we don't need storage hardware?
If IT practitioners are getting confused, what chance does an end-user have. I weighed in to the debate, with a certain amount of trepidation, to try and describe succinctly, in fairly simple, and by necessity, narrow terms a view of Big Software. Here is what I wrote
"I feel like I might be taking my life in my hands by jumping into this debate😊 However,…
Why “data” networking, is a fairly simple answer. In the old days there were analogue voice networks and digital data networks and never the twain met. Clearly that is no longer the case, but old terminologies take a long time to die.
Data is clearly not hardware. Network applications are software, but the rub was that, for example, all storage devices had an inbuilt controller which was a mix of proprietary hardware and software. The software was actually called “firmware” (OK, product marketing speak). In a software defined world the control plane (controller) has been removed from the hardware and defined in software alone in such a way that the same piece of (industry standard) storage hardware can now be configured in many different ways in software without the need for expensive and time consuming firmware upgrades.
This has spread to other areas such as networking where the tasks of proprietary switches and load balancers (with embedded controllers) can now be handled by generic industry standard processors far more flexibly and cost effectively.
Even mainframe operating systems have been developed in such a way as to remove the need for proprietary hardware chips and controllers. The Unisys MCP operating system can be bought and run on standard Intel servers without the need to buy an expensive Clearpath mainframe.
Ultimately you always need some hardware to process stuff, although DNA based storage may somewhat blur even those definitions! The bottom line is that you can use the same basic, industry standard, hardware building blocks which brings cost benefits, and they can be reconfigured using software without the need to program for and then upgrade, specific proprietary controllers bringing benefits of increased velocity and agility to systems development and deployment."
I hope that helps. Some of you will think that Big Software is really another way of saying "software-defined storage, software defined data centre etc. etc. Yes it is. But at Bloor we see Big Software as an over-arching approach to the provision of information systems that helps organisations move beyond digital transformation to become Mutable Businesses...businesses that are able to flex and change all the time as new threats and opportunities appear.