Virtualization - evolution or regression? Strange question. But hasn’t virtualization taken us back to the same principles of the mainframe computing platform I discussed in the early days?
Think about it! Our physical servers are centralized in the datacenter. They have business continuity features like high availability, distributed resource scheduling, and fault tolerance. There are virtual servers, desktops, and application instances running on them to maximize resource capacity. And now companies are starting to introduce thin client desktops with virtual desktop infrastructure (VDI) solutions to help solve globalization and externalization. Hmmm.... Isn’t a thin client the same as a dumb terminal? And isn’t a virtualized ecosystem strikingly similar to the mainframe environment?
Considering "cloud computing" was originally coined in 1966 by Douglas Parkhill's book The Challenge of Computer Utility it shouldn't be surprising that some of the foundation is rooted in mainframe principles. The book describes elastic provisioning, online delivery, and the perception of infinite supply. Sound familiar?
Infrastructure and operations focus now moves from data center consolidation to offering dynamic infrastructure as a commodity to ensure innovation helps meet business goals. This can be clearly seen in IaaS and PaaS strategic initiatives. Basically – IT consumerization. IT as an industry is beginning to clearly define the services it provides to the business, rather than the activities it executes to provide those services. Defining and valuing the services it provides within infrastructure and operation services will help provide IT leaders with a realistic cost of operating the Wintel environment. And with a highly virtualized environment we can provide elastic provisioning, rapid delivery, high business continuity, and the perception of infinite supply.
IDC estimates the worldwide cloud market expenditure will be more then $70 billion by 2015 - That is a 25% annual growth rate.
But this new transformation is a tricky environment to navigate. Especially when all stakeholders representing infrastructure IT, development, and business partners aren't working together to chart the course. There should be a umbrella of options covering all aspects of cloud computing options that provide the business with the best financial benefits, agility, and security. The portfolio should include Private Cloud, Public Cloud, and Hybrid Cloud solutions as well IaaS, SaaS, and PaaS.
One area I find very fascinating is the emergence of internal PaaS. Open source solutions like Cloud Foundry for java and .Net solutions like Apprenda. In a large organization I can see IaaS solutions for consumers that require full virtual machines. These could include infrastructure partners that support Active Directory, IP telephony, and infrastructure support tools. I think in the not to distant future, internal PaaS will be the solution for most corporate web applications. And our business users will use SaaS solutions to select business services.
These services won't just be hosted internally, but will also be extended to hybrid clouds.
Although cloud computing holds several benefits, it isn't a panacea to solve all IT problems, in fact if forethought isn't used for designing a corporate strategy that encompasses all aspects of cloud computing then you could lose significant amounts of investment capital and place your business applications at risk. I plan to discuss this more in an upcoming blog post.