Cloud Computing A Perspective Date:   Thursday , November 05, 2009 Overview Cloud computing as a concept glues together several other, often independent, concepts in a complimentary way to open up new operating models and opportunities for the IT industry. Cloud Computing has evolved from, and extends into, several concepts that have been around for some time, such as SaaS, utility computing, grid computing, virtualization, realtime infrastructure, Web platforms, and SOA. Gartner defines cloud computing as ‘a style of computing where massively scalable IT enabled capabilities are delivered as a service to external customers using Internet technologies’. Cloud computing basically has three layers, and using these layers mainly following services can be provided: 1. Infrastructure as a Service (IaaS) * Physical infrastructure as a service - traditional data center services offered using utility pricing models. * Virtual infrastructure as a service - virtual servers, logical disks, VLAN networks, and so on. 2. Platform as a Service (PaaS) – runs on top of the infrastructure layer * Web based development tools (aka IDE) - could more accurately be referred to as Tools-as-a-Service (TaaS). * A run-time application platform that enables running applications in the cloud, typically on top of an IaaS and delivered as SaaS. 3 Applications or Software as a Service (SaaS) – runs on top of infrastructure and platform * Application components as a service (can be argued that it is part of PaaS) * Software as a Service (SaaS) Looking at the above, one is inclined to agree with Gartner when they describe it as ‘XaaS’ or ‘Everything as a Service’. The cloud based delivery of capabilities demands that a variety of vendors and providers come together in a cohesive manner and, therefore, it is naturally a complex ecosystem. The following diagram, though not comprehensive (in terms of vendor coverage), provides a good overview of the current vendor inter-play: Cloud computing provides several benefits to the consumers: * It allows them to convert their fixed costs related to IT to a variable utility pricing model. * It provides freedeom from bothering about how the IT behind their business works, and allows focusing on just their business. * It allows them the flexibility to scaling their operations (elasticity) with ease, while offering higher reliability and performance. Components of Cloud Computing Like a jigsaw puzzle, delivering applications and services on the cloud requires integration of various pieces. The Cloud Delivery Cycle (CDC) attempts to provide the overall picture of these components working together. Hardware Machine Pool The first key component is the hardware machine pool. It can also be referred to as the ‘Cloud Infrastructure’. This pool could have commodity class machines – racks and racks of them. On the other hand, we could have server class machines and then use virtualization to chop them up into smaller pieces and offer each piece as an independent machine. Application Machine Image (AMI) In order for the machine to be useful, we need a software platform and a business application sitting on it. Application Machine Image (AMI) is the process of bundling of applications in a standardized manner so that it can be deployed on a specific cloud platform. The process of provisioning a machine and deploying an AMI automatically can also be referred to as elasticity. Governance The whole process of provisioning with other business rules can also be termed as ‘cloud governance’. Governance is essential because elasticity has cost and SLA implications. A governance model that will satisfy a diverse set of customers with varied operating parameters. The following two models are only indicative and do not represent the complete suite of possible solutions: *Elasticity Subscription model – Platinum, Gold, or Silver. In this model the aspects like size and revenue of an organization will be immaterial. The elasticity model is associated to the subscription. A platinum subscription will offer customers self-service, customer support, and near realtime on-demand elasticity benefits. This model is targeted for large-scale enterprise customers. *Elasticity Insurance model – In other words, every customer may choose to have entitlements for a set of buffer boxes (predefined upper limit) that can be made available when the customer’s application needs it. The buffer percentage can be derived from the average size of the infrastructure for a customer (subject to the upper limit). This model will be commercially viable for small to medium sized customers. While the models attempt to offer a solution to the problem of governance, the complexity of support systems such as monitoring, triggering, elasticity, and billing will remain or perhaps increase. Another form of governance is around physical hardware. The rigor needed to establish rules around getting the physical machines ready to be provisioned are critical. Integration Gateway Although proponents of cloud visualize everything as moving to the cloud, the reality will probably be a mix of on-premise model extending into the cloud where possible. It will be a judicious amalgamation of commodity services on the cloud and core business processes residing within the enterprise that will yield the best return on IT investments. Every function of IT should not be, or cannot be, put on the cloud and that means that there is a significant opportunity waiting to be leveraged - ‘integrating’ cloud services with in-house IT systems. This is where SOA can play a critical role and it is important that a services based secure handshake is needed between the cloud components of an application and the rest of the application residing in an enterprise. Public or Private? As always, with any technology there are multiple choices, and with cloud computing it is public as against private. There is no silver bullet that can help take a decision, but here are some pointers that can help in arriving at a consensus: * If there is a sizable investment in hardware that is relatively unused, then it can be used as a means to establish your own private cloud to service various business units within the organization. * For a small to medium business, a public cloud is certainly a viable option to manage IT expenses. * Data security can be of paramount concern for many financial institutions; therefore, a private cloud can be the obvious choice. * The other side of public cloud is that the availability of the service, hardware, and software is not directly in control of the organization and hence an element of risk comes into play. * IT can leverage computational power as needed. This dynamic expansion and contraction of the IT infrastructure is perhaps one of the most powerful capabilities of cloud computing if done right. A public cloud will typically deliver this better. Conclusion Cloud computing has a lot of promise as evident from the fact that software powerhouses such as IBM, Google, and Microsoft are investing in it heavily. On the other hand cloud computing, in many ways, seems to be drawing us back to the UNIX philosophy of a ‘central computer’ doing the bulk of the processing. It is said that the reincarnation of technology brings back an idea with certain augmentations. Is it true for cloud computing as well? Will it deliver the goods it promises to? Only time will tell. The author is Director of Solutions Development, Technology Innovation Center, Perot Systems.