In 1991, I was assigned a team at PwC to build something called “2O to 3O”. Dun and Bradstreet Software, especially its McCormack and Dodge (M&D) unit, was an important practice area for PwC. M&D had released a new General Ledger 3.0, a product with a multi-dimensional code block and a number of other new features. They did not, however offer 2.X customers any upgrade software to get to 3.0. The thinking was the new product was so different, customers would re-implement. At PwC, we sensed an opportunity to develop a mostly automated solution - we ended up with a PC based configurator, a set of mainframe migration programs and a consulting package.
My team was surrounded by cynicism. Feedback from M&D was it would be too difficult, we were wasting our time. Feedback within PwC was why are we cannibalizing our service opportunity? Anyways we got it done (the master and transaction files were relatively straightforward, the control file where you stored allocation, report writer and other specs far more complex) but in the mean time, the market had moved. Clients did not want to migrate to another mainframe product. Indeed D&B itself has announced its Smartstream client/server product and that froze customer interest in 3.0. Ironically, having shown we could build a migration product, D&B called us a short while later and asked us to help them build migration tools to Smartstream. That is another story but I have always wondered why software vendors do not think of automated migrations when they design next-gen products.
So it was nice to see Larry Ellison present last week a modern day version of an automated upgrade that my team had developed. Its called Oracle Soar (to the Cloud) and the components are in the graph below.
The Soar functional scope is much wider than the General Ledger we focused on, it is going from a client/server to a cloud platform, and there are lots more client/server customizations to sunset – so it is several times more complex.
And I love the marketing pitch – it is the “last” upgrade customers will need to spend much time on. Later upgrades in cloud settings are handled by Oracle, in background.
My questions
- why did it take Oracle so long to develop this?
- why are other vendors not writing these for their products, and even those around legacy products of competitors?
- why are systems integrators, while bragging about automation projects at customers, not building these?
I know the answer to the last one. The SaaS model has shrunk hosting, application management and upgrade services that were lucrative in on-premise world. Implementations remain labor intensive and are good for business if you want to continue to sell rate-card based services.
The reality, as I showed in multiple case studies in Silicon Collar, is automation makes your workers smarter, speedier and safer. Automated migrations are clearly speedier. They also allow your consultants to focus on smarter, more creative work. And they are safer – at least in that they give you an advantage over competition which continues with hand to hand combat.
So, I hope other software vendors and outsourcers follow the Oracle lead here.
BTW, spend the 40 minutes Larry Ellison, Beth Boettcher and Steve Miranda spend discussing Soar – worth your time even if you are not an Oracle customer.
