Open any respectable publication or market report today and you cannot escape the mention of “Cloud Computing”
(See you got it here too !).
The media all around us is so full of this buzz word on the Gartner hype cycle, “Cloud Computing”. With so many different messages and definitions, the cacophony of different interpretations of Cloud Computing only serves to confuse most people. There are many new strange looking acronyms like SaaS, PaaS, IaaS, StaaS, etc running all over the place adding to the chaos. This article is an attempt to “demystify” Cloud Computing, hopefully without adding to the confusion we already have around us.
So what is Cloud Computing all about?
Gartner defines Cloud Computing as “A style of computing where scalable and elastic IT-enabled capabilities are delivered as a service to customers using Internet technologies.”
Forrester defines Cloud Computing as "A pool of highly scalable and managed compute infrastructure capable of hosting end-customer applications and billed by consumption."
NIST (National Institute of Standards & Technology) has released a detailed, “Official” version of the definition of Cloud Computing running into several pages.
I have not yet started with definition of Cloud Computing by multitude of other “experts” in the market.
In simple terms, Cloud is a “disaggregated mass of storage, network and computing resources accessible in a dynamically scalable form in on-demand mode” … getting more complicated?
By resources, I am referring not to IT trained people as is normally done in IT organizations, but the various devices and applications and networks used by the IT organization and enterprise users.
Types of Clouds
To simplify it further, every enterprise IT system is made up of devices & applications that allow users to store information, share information through networks and compute or process information. Usually these resources are spread across the organization with each user able to use only a fraction of the resources available to him/her. This amounts to significant over-investment and under-utilization of IT resources. For example, most of us use only a small part of all the applications available to us in Microsoft Office in our routine, though we have to pay for the full package. Some people use Excel more than anything else and some others use Powerpoint and some others use MS Word more than anything else, most of the time. Very few users actually use all the applications that we pay for when we buy the MS Office license, with significant enough frequency to justify over even 50-60% of the license cost.
In cloud computing these distributed resources are connected to a common platform that can be accessed by all users across the organization on need basis, leading to ready availability and dynamic scalability owing to tapping the under-utilized resources to their fullest potential. This is called virtualization of resources – making the functionalities available in the virtual (online) world.
With this kind of a set-up, the organization enjoys dynamic upward and downward scalability, high resilience and flexibility and better utilization of IT resources. All of this, when done within the limits of an enterprise, is known as a “Private Cloud” or “Enterprise Cloud”. In this case, various parts of the enterprise use the common infrastructure (Cloud) in a pay per use mode through a system of “internal charge-backs”
When done on a larger scale for public access, it becomes a “Public Cloud” of which the Google Documents facility is an example. Here the revenue model is either pay-per-use based or free to use to grow volumes and earn revenues to advertisements.
A third kind of Cloud is a Hybrid Cloud set up where there is a secure private cloud on the enterprise side which is connected to a public cloud environment in a secure and controlled manner to enable access to information and services to customers.
Cloud Enabled Services & Cloud Enablement Services
Various service providers offer a variety of services leveraging the Cloud, based on their own capabilities and investments.
Application vendors offer SaaS or Software as a Service which is a refined version of the erstwhile hosted application services in the name of ASPs. In SaaS the applications are available on a pay-per-use basis and is also referred to as “Apps on Taps” by some.
StaaS or Storage as a Service is offered by providers who set up huge storage farms (yes they call them farms) and make storage available on a pay per use model such as pay-per-GB or pay-per-month for unlimited GB, and many other pricing models.
IaaS or Infrastructure as a Service is offered by providers who put together storage, network and computing infrastructure on a large scale and allow users to use it on a pay-per-use basis. They use various models such as server rentals per day or pay per computing cycle and many other such pricing models. Bandwidth allocation and server resources are essential elements of IaaS. The IaaS architecture specifies dynamic scaling of bandwidth. A cloud computing installation should never be overwhelmed by peak demand, because the infrastructure service should be able to quickly add the resources needed to accommodate peak demand.
PaaS or Platform as a Service is the delivery of a computing platform in a pay-per-use service mode. The service delivery model allows the customer to rent virtualized servers and associated services for running existing applications or developing and testing new ones. Enterprise IT managers often prefer to use PaaS for setting up application development and testing environments in an upward and more importantly, downward scalable mode, to avoid sinking huge infrastructure investments in resources that are required to full capacity for short time intervals. Many small IT service providers and ISVs (independent software vendors) also find PaaS very useful.
Some IT service providers also offer Cloud Enablement Services and Cloud Testing services.
Cloud testing services refer to third-party testing of a Cloud set up by someone else to check for compliance with the mutually agreed upon expectations as per the terms of the contract, before payment is released to the provider who created the Cloud.
Cloud Enablement Services refer to services aimed at helping enterprises set up their own enterprise Cloud. This is a combination of a series of services as listed below:
Cloud Readiness Assessment:
Consultants from the provider organization study the existing IT set up and test it for readiness to move to a Cloud environment. This involves assessment of the portfolio of applications, hardware and connectivity infrastructure and also the way the processes are organized across the organization. Based on this assessment, the Consultant creates a roadmap.
Running the roadmap:
In this phase, the provider focuses on optimizing the IT resources through various steps such as applications portfolio rationalization, process re-engineering for optimization, geographical consolidation, etc. Once the existing IT environment is fully optimized in its current offline form, the consultants work on virtualizing it and creating relevant charge-back mechanisms to meter and charge for usage.
Testing the Cloud:
The Cloud that is ready in the background is then tested offload and on peak load to ensure it lives up to its promise of dynamic scalability, resilience and functionality.
Launching the Cloud:
Once the entire Storage, Network & Computing resources portfolio is virtualized and is ready for pay-per-use in an on-demand mode, training is conducted for the IT staff to get them ready to manage and maintain the Cloud. The Cloud is then officially launched by migration of the access from offline to Cloud mode. In many cases, the entire process is transparent to the end users and they may not even know their access has been transferred to Cloud mode, unless required by specific regulations.
Benefits of Cloud Computing
The way Cloud is created, or meant to be created, it has the following inherent benefits for the user:
Elasticity / Dynamic Scalability:
Just like a rubber-band, that can be stretched as much & when required and post use can return to its original steady state, Cloud offers dynamic upward and downward scalability where additional capacity can be added ON DEMAND in a matter of minutes or even seconds, at times even transparent to the users. Scale down can also happen equally fast when the need for excess capacity does not exist anymore. This is very useful for known cyclical spikes and also unknown volume spikes in activity. A good example is an online retail store that has a huge cyclical / seasonal volume spike around say Christmas where the retailer would have to provision infrastructure to handle volumes as high as twice or thrice of the usual steady-state volume. If the infrastructure is unable to scale up when required, the site would crash and there would be a huge opportunity loss. If the retailer prepares for the spike and provisions for the peak infrastructure all year round, it is a waste of investment for the rest of the year. It is much easier to provision for the additional infrastructure through Cloud service providers, for only the specific time of the year. The provisioning is dynamic in a matter of seconds or minutes as against waiting for weeks to install an additional server. Another example could be that of a portal of an anti-virus software vendor that faces the peril of sudden, unpredictable rise in traffic on a virus outbreak, or an online stock trading portal that may face sudden spikes in volume on major positive or negative news.
Conversion of IT CAPEX to IT OPEX:
The cloud offers an effective way of converting IT CAPEX (capital expenditure) to OPEX (operational expenditure) with a pay-on-use pricing model. It negates the need for forecast-based sunk investments in IT infrastructure too.
Speed of Provisioning:
IT managers can provision for a specific development or testing environment within minutes if not hours instead of the weeks that are required in the offline world. With the scalability being dynamic and various pricing models available, Cloud offers a speedy way of provisioning IT resources. Cloud also helps reduce time for application roll-outs and upgrades for months and quarters to a matter of hours.
Disaster Recovery & Business Continuity:
With shared IT infrastructure, often physically located over a wide geography and with provision for dynamic mirroring, the Cloud can provide a virtually infallible IT infrastructure.
Especially in case of enterprise Clouds, underutilized infrastructure can be used much more effectively in a shared manner. For public and hybrid Clouds too, the long-term cost of pay-per-use is much lower than the long-term cost of ownership due to high utilization. Maintenance and upgrade costs are reduced to almost negligible levels through SaaS and virtualized infrastructure.
Challenges & Concerns
Now that we have looked at the silver lining of the Cloud, let us also look at the dark side, as lightning bolts usually come from there and could have lethal consequences. Despite its obvious (aren’t they very obvious now?) advantages, there are certain challenges and concerns too.
There is always a security threat in using shared resources. However this can be addressed with enterprise Cloud with a thick and smart security layer around the infrastructure and putting together specific identity and access management policies and procedures to minimize the risk. In case of public Cloud, especially in B2B scenario, appropriate vendor evaluation and strict SLAs and damage clauses could protect your interests. Discretion MUST off course be used in deciding what goes on the public cloud. Risk can also be managed by using a selective Cloud migration strategy, instead of going all-out on the Cloud.
With Cloud, having a disaggregated mass of IT infrastructure and applications, often spread across geographies, there are concerns about where the data is being stored and regulations in some geographies may prohibit certain kind of data to be sent out to less secure locations. There is often no way around regulations and hence these regulations could be used as one of the guiding principles in deciding what goes on the Cloud and what does not.
Not all software vendors are yet offering SaaS or Cloud licenses for their products. This becomes a challenge when creating a Private Cloud within the enterprise where one of the features will be “apps on taps”. Decision to enable SaaS on enterprise Cloud ought to be made after discussing these issues with the relevant vendors and arriving at some mutually agreeable solution.
Operating a Cloud enabled environment requires HUGE bandwidth availability with suitable redundancies to avoid crash of business operations in case of connectivity outage. Some calculations ought to be made to check the potential increase in bandwidth cost against the expected benefits accruing from Cloud.
Most organizations would struggle to find a single Cloud provider that can address all its requirements. This means signing up with different vendors and managing the multiple relationships, which often has the potential to sap the management bandwidth significantly. There is also the concern around interoperability of offerings of different vendors. This risk can be mitigated through systematic vendor evaluation and selection and with well drafted SLAs.
There may be some more challenges and concerns potential buyers and actual users of Cloud may be facing, but the above are the most common. Most of these can be addressed with a little discretion and proper planning.
Like any new technology or concept, Cloud computing too has its enthusiasts and naysayers. However it is obvious that Cloud is emerging as the new industry standard for organizations globally due to the obvious benefits it offers. As it happens with any new concept or technology, early adopters will adopt for competitive advantage and laggards will follow over time when the benefits are proven and the creases are ironed out. As providers and buyers mature and as the hype settles, Cloud computing will over time become just another way of using IT resources in a shared manner. Till the time that it actually happens, it will be good to remember the age-old adage –
“Look Before You Leap”.
Ravindra Shriram Datar
Head – Global Marketing
Datamatics Global Services Limited