Companies typically spend about 10-15% of their annual revenue on IT. That would be 525 to 768 Billion USD in 5 US states– NY, NJ, CT, CA, TX alone! By many accounts “cloud” promises to save 10-20% of this cost by enabling inter and intra organization sharing of compute infrastructures. Not to mention equivalent savings opportunities from exploitation of additional avenues of IT optimization. For the same 5 states above, the savings would translate to more than 52.5 to 105 billion dollars annually. Sounds really good—so why are many companies not rushing to rapid cloud adoption and IT optimization? Two primary reasons– concerns about risk and lack of proper motivation for key players in the IT eco-system. Risk we could potentially mitigate, but for cloud and IT optimization to really take off— proper incentives must be in place for all the participants in IT.
Significant improvement opportunities do exist The avenues for optimization embrace the same age-old concepts— maximize utilization and eliminate unnecessary work.
Maximizing utilization
Virtualization and shared clusters are two avenues for increasing utilization each with its own pros and cons. In the case of virtualization allocation of work among the shared resources are managed by the hypervisor while in clustered systems balanced sharing is achieved by appropriate distribution of transactions across the underlying resources. In both cases it helps if application design is amenable to scale-out. Virtualization and clustering approaches impact application performance differently and have different cost structures for support. Application grouping and placement are important considerations to ensure peaks and valleys of individual applications properly overlap so that there are no radical peaks in demand at the aggregate level.
Reducing unnecessary work
Typically this is where we would find the most improvement opportunities. Often precious resources are kept busy with work that can or could have been avoided. This happens when the “do once share by many” principle is not fully exploited. For example, an application makes the same service call to get the same information multiple times during the process flow of the application. Another typical example would be when a computation is done at the wrong tier of the infrastructure causing unnecessary traversal of many compute resources. We have also seen examples where, in the name of standardization, the same priorities and polices are applied uniformly across applications that generate a billion dollars in revenue and ones that generate only thousands. In such cases, proper application re-architecture and adjustment of priorities and policies not only helps application performance and availability, it lowers demand on resources as well.
Streamlining processes and automation
How many weeks and months of preparation and planning does it take to deliver incremental business function? How many people are involved? And after all that planning and preparation do releases meet production expectations? Not always! Organizations need to implement repeatable processes with predictable results. And if one can achieve that, the processes themselves can be automated. End-to-end automation is another great savings opportunity— but can only be exploited if technology diversity is managed and standards are implemented.
Standardization
Typically large organizations will have many similar IT products each stated to have that little something different or extra. Entrenched diversity in technology adds to support overhead and is an impediment to shared infrastructures and IT transformation. Cloud adoption naturally abstracts out applications from the underlying infrastructure enabling de-facto standards– which in turn leads to simplification and reduction in diversity of IT footprint.
While opportunity for significant IT improvement do exist in most organizations– the barriers to change are significant as well. The most common one is the fear of the unknown or risk.
Risk concerns act as barriers to cloud adoption
Cloud security and having core business process logic and data in a public infrastructure that is controlled by another business entity is a key concern for many organizations. Isolation is another—companies are concerned that another company’s rogue application could impact their core applications. Vendor or provider maturity is another area of concern—businesses are not willing to place significant bets unless the technology and vendor viability has been solidly proven. Often, in large organizations when faced with complex production issues the application team and infrastructure team will point to each other. Sound familiar? Problem resolution and accountability will become complex inter-company issues with cloud adoption. Unless metrics and tools are able to provide much deeper visibility and control into the service provider infrastructure, we anticipate organizations can be exposed to revenue drain from potential flaws with charges and charge models.
Cultural barriers to cloud adoption
The key stakeholders of IT in an organization happens to be the CEO/CFO/CIO, the many Vice Presidents of IT, the manager and the individual contributors. And let us not forget the big IT vendors who are perhaps the most important players in the IT eco-system. What incentives to these participants have in enabling significant cost-savings for the business?
Big vendors will not cannibalize their own revenue streams for the benefit of their clients
Imagine a company with north of 1B dollar IT budget—chances are there is a big vendor that takes 20% of that pie or roughly 200M dollars. A 20% cost savings over the entire spend for that organization could translate to a 40M dollars loss in revenue for that vendor! Unfortunately, most large corporations will also rely on the opinion of these big vendors before they take a step forward and therein the chicken and egg problem. There are a few resolutions to this barrier—1) New vendors come in and cause a mix-shift in the eco-system first with adoption in the small business space leading to penetration into the enterprise space 2) Big IT vendors reinvent, adjust and grow revenue in new opportunities so that they are able to jettison the old and 3) corporations seed their organization with a culture of leading innovation.
Cloud value proposition for CEO must be one of growth and competitive advantage and not just cost savings
In preparing for every quarterly report CFOs are under a lot of pressure to report improvements to the bottom-line. On the other hand no CEO can ever say that their singular goal is one of reducing cost—the best way to eliminate cost completely is not to have a business! CEOs focus on growing the business and increasing shareholder value. Margin improvement with modest client acquisition and significant client retention is a viable business model for many corporations and IT cost savings is an important consideration for these companies. However, in businesses that can have large revenue fluctuations (an example would be currency fluctuations that cause 1B+ dollar revenue swing in a large bank) a 20% savings on 10% of the revenue stream is only a small consideration, relatively speaking. For CEOs of these corporations the more appealing value proposition of cloud and optimized IT is that cloud can lower time to value and is critical to creating an environment of technology driven business innovation in their organization.
CIO/VP of IT must be compensated for value not budget or size of organization
CIOs are accountable for a key enabling business function but they do not directly drive line of business revenue streams. And yes, as technology continues to play an increasing role in business growth the modern CIO has more equitable standing than before on the C-suite ledge. But the size of the organization and the budget managed continues to be the yardstick of power and influence for many CIOs. Lets face it, a CIO with a 1B dollar budget would not have the same bragging rights and industry influence if he or she were to have a budget of mere 100M dollars. Some corporations are addressing this by expanding CIO roles to include core business functions such as product management. Another approach is to make senior IT roles to be term based and compensation levels defined by value delivered- not size of organization. Of course, for all of this to work effectively, organizations will have to define key metrics that objectively determine value. A successful transformative leader in IT can potentially be a successful transformative leader in other parts of the business and rotational assignments constitutes a good way to seed IT savvy business leaders across all business functions of the organization.
In conclusion
There is more to cloud than just virtualization and shared infrastructures. The sharing can be at many different levels—some constitute easy wins while others may require more savvy. The potential for gains and the choices depend on current state and the desired end-state, which may differ from business to business. Often, folks who do not welcome the change will throw down the gauntlet of “risk” and senior non-IT leaders who desperately want the transformation, however, cannot assume accountability for something that they know very little about, and will cave in. Having metrics for everything- cost, availability, performance and risk will help in this regard. Ability to produce objective measurements and metrics quantifying “before and after” will be key to success of a company’s IT transformation initiatives.