A Non-Technical Overview of Cloud Computing

by Scott Trathen, Ph.D.

Cloud computing is currently the most over-hyped and least understood technological innovation since the advent of the Internet. The latter tremendously impacted businesses of all sizes globally, and the former has the potential to similarly and dramatically transform the business and economics of IT over the next few years. The purpose of this Strategy Brief is to expose a business leader to a comprehensive and yet simplified overview of the potential business value derived by the improved technology delivery mechanism better known as “cloud computing”. The key takeaway is there are many choices that will prescribe your journey towards the cloud – the least risk is to bet on the most comprehensive management capabilities.

What is Cloud Computing?

Cloud Computing is a way of providing and consuming IT Services. Many in the industry are coalescing on the definition provided by the National Institute of Standards and Technology (NIST) in the US which gives the most widely recognized definition of the attributes (elastic, on-demand, pooled resources and metered services), service models (Infrastructure as a Service (Iaas), Platform as a Service (Paas), Software as a Service (Saas)), and deployment approaches (Private, Hybrid, Public) comprising cloud computing. Most of the strategic IT vendors are aligning their messaging to these standard terminologies, and developing offerings in hopes of securing a stronghold in this nascent marketplace projected to be worth nearly $150B by 2014 (Gartner).

At its essence, cloud computing is the aggregation of resources (compute, storage and network) so that capacity requirements can be dynamically modified to match fluctuation in system workloads. This resource matching can be automated, or self-provisioned by line of business (LoB) resources interacting with the cloud environment in business terms (# of transactions, concurrent users, time to result). The value proposition of cloud computing also includes agile compute infrastructure provisioning (“purposing”); or nearly instantaneous resource availability (growth, new application, disaster tolerance). Using these same principles, cloud computing can readily recycle and repurpose resources into higher value computing needs. Inherent value is derived from oversubscribing the resources in an intelligent and optimized fashion, thus creating increased economic leverage. Like the typical airlines approach to overbooking flights, this transformation in “purposing” of IT resources maximizes the capabilities of the asset, and hence the return on the investment.

Regardless of the cloud computing deployment approach, oversubscription and dynamic resource matching require a comprehensive portfolio of security and management capabilities. It is these capabilities that primarily differentiate cloud computing from virtualized datacenters and allow for a more fully integrated value chain within IT.

The Cloud Computing Continuum

The greatest economic advantage to be gained is through the Public Cloud - wherein economies of scale are leveraged along the efficient frontier. Massive, centralized data centers allow for cost sharing arrangements of expensive capital prerequisites (purchasing power, real estate, HVAC, power, network, compute, storage and labor) across organizations with a subscription-based operating cost model. These cost sharing arrangements smooth the cost curves for incremental units of compute, and lower the carrying costs associated with compute inventory supporting “limitless scalability”. Additionally, data center infrastructure is capital-intensive and expensive to operate. The financial benefits of aligning computing requirements with an operational cost model are significant and merit consideration on their own accord. From a technological perspective, these cost sharing arrangements can introduce technical tradeoffs in the form of security, privacy and compliance concerns. It is incumbent on the industry and IT vendors to innovate against these concerns as the model matures, for the full economic potential of cloud computing to be recognized. In the meantime, these economic benefits can be realized situationally where the cost or agility benefits outweigh and mitigate the identified concerns. A few key examples are test and development environments, definitively cyclical/seasonal computing needs and development cycles where time to market/revenue are clear.

Moreover, organizations seek to replicate a subset of the economic benefits on a smaller scale through Private Clouds as in many cases they are perceived barriers to public cloud usage. In this case, the cost sharing arrangements occur within an organization without the concerns introduced by sharing compute resources with other entities. For these reasons, Gartner reinforces this adoption approach, “Through 2012, Global 1000 IT organizations will spend more money building private cloud-computing services than on offerings from public cloud-computing service providers.” They continue by projecting “two-thirds of large enterprises plan to pursue a private cloud computing strategy by 2014.”

When it comes to Private Clouds, many of the vendors in the marketplace continue to extoll the virtues of virtualization. However, IT organizations don’t exist to deploy and manage infrastructure – they deploy and manage applications delivering business value that require infrastructure. Holistically considering the purpose of IT, it’s not about maximizing virtual machines, rather it’s about the applications that run on them!

Properly architected Private Clouds allow a business and IT organization to manage the compute components (server, storage, network) from the application layer. In doing so, the entirety of the service components can be managed as an application service. This provides the opportunity for the application service (end-end) to be managed to a Service Level Agreement (SLA). Thus the business/business owners are empowered to make decisions (availability, scalability, disaster recovery) with full transparency to the associated costs. Inherently, these decisions are enabled by a combination of people, process and technology. The merits of a properly architected Private Cloud automate the latter two, and maximize people/labor in the process.

Managing to an SLA suggests the cultural transformation of managing to a service level – not managing the individual physical /virtual machines or components. A defining characteristic of Private Cloud is the criticality of the management infrastructure which surfaces of premium importance, especially when considering that most existing IT infrastructures are heterogeneous in nature. According to Morgan Stanley, “less than 10% of companies have adequate management tools for their virtual environments.” Aside from gaining IT labor leverage, this gap ensures a Private Cloud returns are capped at the virtualized component level. Moreover, the innovation in management capabilities to assist the advent of Private Cloud is significant. Gone are the days IT needs to manage the individual datacenter complexities, rather a holistic view of the application service levels provides automated, dynamic and proactive management of application templates and compute profiles.

Gartner estimates the cost of a traditionally continually available IT service (one site) to be nearly 6.5X the cost of a standard IT service. Trends indicate the expectations and reliance on computing by end users/consumers continue to increase, driven by unique application providers. As this reliance transcends the enterprise, end users continue to expect their mission critical applications to always work. Yet, the cost of enabling such a myriad of application to be available is exorbitant using traditional means. This is the transforming vision of the Cloud, and the reason implementing Private Clouds is gaining such traction in the marketplace.

Consider the merits of a computing environment monitoring the health of all the elements in the integrated value chain, and proactively migrating away from degraded performance to avoid downtime for critical systems. Or, optimizing the compute needs via moving averages – without incurring the minimal return of dormant assets scaled for peak demand. Or, delivering value to the business by using idle compute resources for strategic contingency modeling outcomes (What If analysis, Monte Carlo simulations, Game Theory). The ability to optimally and dynamically repurpose computing assets towards high business value opportunities such as these mentioned is one of the value drivers of cloud computing.

Along the Cloud Computing Continuum, Private Clouds are an entry point. The concept of continuum suggests an evolutionary progression, and for Private Clouds this equates to a Hybrid model. The Hybrid model is one in which some elements of a Private Cloud are merged with some elements of a Public Cloud. An example is the need for burstable computing capacity to accommodate a spike (foreseen or unforeseen). The Hybrid model is the economic and value multiplier of the Private Cloud, and perhaps the reason for implementing a Private Cloud in the first place. Such an implementation necessitates an advanced and cloud-aware orchestration capability built on rules and policies for leveraging such extensibility of the computing environment.

Why Microsoft Corporation?

Leveraging the transformational value of Cloud Computing is not as simple as buying a product, or flipping a switch. It is a journey, across a period of time and cycles of innovation – in the marketplace and within an organization’s culture. A baseline for this journey is a good understanding of the IT technological and operational capability requirements and their maturity levels. The gap between this and a desired end-state allows for a path to be plotted for the journey. Projections indicate organizations are embarking on this path now and in the near future, and so critical relationships are being formed that predict the odds of success. Microsoft has a vested interest of helping its vast customer base succeed in realizing the benefits of this transformation and is the only vendor capable of offering solutions across all service and deployment models.

Microsoft has invested $2.4B to locate datacenters across the world, and even more substantially to ready the portfolio of software assets for cloud computing. This includes offerings across the Cloud Continuum (Iaas – SaaS) with HyperV Clouds, Windows Azure and Office365 & CRM Online. Uniquely, Microsoft offers the underlying infrastructure (Management, Operating Systems, Databases) and the end user applications (CRM, Exchange, SharePoint) and can best optimize the integrated application stack for cloud computing. With such a motivation, and a solid financial backing to execute on the necessary investments, Microsoft stands alone in the marketplace offering customers the flexibility to adopt the cloud on their terms.

Now What?

The purpose of this article is to expose a business leader to a comprehensive and yet simplified overview of the potential business value derived by the improved technology delivery mechanism better known as “cloud computing”. If successful, the question becomes now what?

Much like any other business decision, the first step is to determine the viability of a business case for cloud computing. Do the benefits of Public Cloud outweigh the tradeoffs for your organization, or a subset of your computing needs? Can you gain economic advantage and increase computing availability by leveraging a Private Cloud? In either case, strategically identifying the right application or service is critical. Start small, understand the selection criteria, and seek professional help for a Proof of Concept. This is a transformational journey for IT, and successes, failures and lessons learned on a small scale now are worth their weight in gold.

A summary of key steps is as follows:

• Build a business case by understanding which services have potential for one of the Cloud service and deployment models from a cost savings, business agiliy or service quality perspective. This means building an assessment strategy for the portfolio of services based on appropriate criteria.
• Many applications will take time to transition to Public Cloud based delivery, therefore, start building a Private Cloud strategy if not already in place.
• Ensure a broad architectural function (such as an Enterprise Architecture team) are engaged to help map business need to technology solutions and start building a roadmap for change.
• Build a risk management approach for identifying mitigation strategies for the implications of service and deployment models used.
• From a shortlist of candidate services for Cloud deployment, identify low-hanging fruit that provides the simplest deployment process and yet delivers business value in a measurable way.
• Determine the extent of any gap between IT capabilities expected for Cloud service models and your current IT capabilities.
• Leveraging and delivering Cloud services is not just a technology oriented solution, therefore you must determine process capability and maturity of your organization and contrast that against what may be expected for Cloud service models.
• If considering adopting a Cloud application platform, contrast current application development strategy and approach against that defined by the platform provider.
• Plan a pilot for that initial "low hanging-fruit" project identified and use the new service to not only understand benefits derived but also confirm technological, operational and organizational impact.
• Of course, no matter which Cloud service or deployment model is considered attractive for pilot, Microsoft can help with everything required to get the pilot implemented and ensure success