Because of costs, complexity, lack of resources, and risks - higher education institutions in Philadelphia including University of Pennsylvania, Community College of Philadelphia, Temple University and Drexel University banded together and setup a non-profit called UNICOL to share an IBM 360 mainframe and all the resources to operate the environment. Users of UNICOL signed up for accounts, and paid for CPU and Disk usage monthly much like I pay for water and gas used at my house on-demand today.
My first company(2) called Phaedra, became a user of UNICOL developing batch based services. In those early days of time sharing we did not call it on-demand, cloud or virtualization. Phaedra was able to prosper selling IT services to organizations that had high volume, repetitive processes who desired automation. Many organizations could not afford to buy mainframes or allocate the physical space,specialized cooling and raised floors to hide all the cabling. Phaedra was able to offer affordable batch computer services– like scoring a final exam for class of 50 students for less than $25. No one could beat that pricing – the test scoring service saved faculty hours of effort and reduced the probability of institutions developing their own test scoring application because it was so inexpensive and provided a nice value.
High education was ripe for re-engineering and computing back in the 70's. Functions were scattered across departments. Many applications included paper based forms. Segregation of duties and control required extra steps. High volume during peak periods complicated staffing. At the same time, custom software was very expensive to setup and sustain. Corporations, universities, government agencies, and hospitals employed hundreds of programmers each to develop and maintain custom software applications like payroll, billing, accounts receivable, registration, accounts payables and so many other functional areas. Time sharing alternatives offering CPU time, disk space, and applications like Payroll popped up all over the region. As computing power became more affordable, more and more organizations established on-premise IT shops, acquired hardware and enterprise software. The pendulum swing justified localized investment rooted in differentiation and control.
Fast forward almost four decades. Virtualization, Cloud computing – and web services is changing the way we acquire and utilize computing power - or at least supplement what we have. Apple, Microsoft, HP, Dell, Yahoo, Google and so many other companies are integrating Cloud computing in their product lines and business models. They are the new digital utilities. Cloud computing examples can be found are all over the web. Open any browser or smartphone and tap the power of applications from Maps to finding an open table for dinner or ball game. We are drawn to the services by convenience and utility.
Virtualizing on-premise servers in the Cloud has gained adoption across many industry sectors. Companies like Amazon, offering an online retail marketplace worldwide - also provides Cloud based virtualized services for IT shops looking for alternatives to on-premise computing. Microsoft, Dell, IBM, HP and others offer private Clouds or online environments segregated by virtual boundaries.
The big gap in Cloud computing is still the lack of shared services starting with identity management. There are so many credentials created by Cloud applications because there is little emphasis on a shared model. Thus, services require the duplication of characteristics and ID proofing - while trying to maintain privacy and consumer control. How many online ID's do you have?
Bridging applications in the Cloud is still a challenge. You can find a large array of Cloud based applications like scheduling a meeting, sharing photos, hosting a conference call, publishing a market survey to collecting data using web forms - yet they are all independent lacking a shared security or data exchange model. These online services represent a new breed of solutions following the utility model, offering consumer’s incremental value based pricing.
In summary, today's digital utilities offers automated services – packaged as transactions, economically justified by volume and sustained by adoption. Digital utilities are an alternative to the on-premise or self managed IT environments duplicating and embedding applications managed by local policies, rules and resources. The trend to outsource and leverage external IT resources will continue to power administrative and academic computing.These are not new concepts in my view. History repeats itself – as do the lessons learned. I hope the sequel is better the second time around.
[1] http://en.wikipedia.org/wiki/Moore's_law
[2] Phaedra, Inc.1976, Pennsylvania Corporation
No comments:
Post a Comment