(Unedited interview with an interviewer from 9.9media India in Oct 2011. It has since been published by them in their monthly publication ITNext.)
What do customers understand from the term cloud computing?
Cloud computing brings in a very different pattern of IT infrastructure deployment compared to what we have seen in the last 15-20 years. Cloud computing enables traditional users of IT systems to access their services from a low foot print device like a laptop or a phone without caring for the details of the computers and programs that may be in use at the data centre or should we say cloud centre. It leads to a lot of simplification and lowers the entry barrier for accessing quality services for the end user. From the application service providers’ perspective, they need not worry about the IT infrastructure details and their management. Availability of scalable, secure and dependable IT infrastructure from a new class of providers called Cloud Service Provider (CSP) ensures that a typical IT application service provider concentrates only on his applications with very little Capital Expense (CAPEX) commitment.
Can you list out certain compelling reasons why enterprises need to be on the Cloud?
Compelling reasons to consider a cloud based solution are many. Reducing Cost and Complexity along with improvement in Scalability, Ease-of-access and stricter centralized control can be cited as top reasons to consider a cloud based model. Cloud model allows the elasticity in demand cope with the elasticity in supply – that is, most of the demand side spikes (both high and low) are taken care of. It results in a situation where you pay for what you actually use. From a more long term perspective, cloud computing pushes the idea of specialization even further by decoupling the IT infrastructure and provisioning issues from the core application issues. For an application owner, this leads to a faster time to market and reduced cost.
Which industry verticals according to you would look at cloud in a serious manner?
Industries that use a lot of distributed processing involving large volumes of data are more likely to consider a cloud based solution. At this point, almost all industry verticals seem to be cautiously optimistic about the potential of cloud. However, problems in integrating legacy systems and not-so-appealing-perception about security in cloud are major inhibitors. So, verticals known to be relatively soft on these considerations are more likely to embrace the cloud faster. Going by this logic, organizations dealing with Education, Entertainment and Unified Communication services should look at cloud more seriously.
Which are the applications that are hosted on the cloud based delivery model? Please specify.
We may categorize applications to two types. One is the application hosting software products that ISVs like Oracle, IBM, SAP and Microsoft make. The other application types are the ones created by organizations by using middleware products from established vendors typically using Java or .NET. At this point, the compatibility between traditional and cloud based middleware options may not be so mature. Even though, Microsoft Azure, one of the more complete cloud platforms, claims to make the transition to cloud comparatively easy, everyone does not seem to agree. It will be sometime before large scale migration of end-user business applications will move to cloud. However, we should recognize that many ISV product vendors are already selling their flagship products for use in a Software-as-Service (SaS) model. Some examples are Oracle DB, SAP, MS SQL Server and MS Dynamics. Salesforce.com is already a successful cloud platform. It has been popular among users needing near-zero investment in sales related activities.
Please elaborate your thoughts on initiatives to move Business Intelligence and Business Analysis applications to cloud?
Use of Business Intelligence has been comparatively less pronounced in smaller organizations, primarily due to the non-linear resource hunger of these applications. In many cases, the cost of using BI is unacceptable due to ROI considerations. However, the equation seems to tilt more favorably when one considers the OPEX model of expenditure used in a cloud setup. Organizations can use BI engines that may be available in a cloud somewhere and hook the output to their applications with a small incremental cost and effort. Therefore, in near future, as cloud based computing becomes popular we are likely to see a more pronounced use of analytics and intelligence in line-of-business applications.
Which of the business critical applications are moving to cloud?
Talking of business critical applications that are meant to run almost all functions in an enterprise, one must mention the dominant position of SAP and Oracle. Both of them have announced the availability of their products for cloud ready deployments. Interestingly, we have also seen Indian companies like Ramco announcing a cloud ready ERP solution. Obviously all these companies see a demand and therefore it is safe to conclude that most application owners are seeing a reason to make their business critical applications cloud ready.
Can you elaborate on the cost structure and billing structure involved in the cloud model?
Cost structure in cloud is essentially based on OPEX (Operational Expense) model. This shift away from CAPEX (Capital Expense) model, where the application service had to provision for expensive IT and networking infrastructure upfront, is a welcome change. This reduces the entry barrier for owners of applications to go live much faster by leasing infrastructure from Cloud Service Providers (CSP) and benefit from the elastic model of pay-per-use. For a CSP, shared use of IT resources by larger number of users leads to higher utilization of the resources and higher revenue. There are already many CSPs who are charging their customers based on the exact use of resources like CPU, Memory, Storage and network. Customers tend to like the more granular billing structure that is based on metered usage of resources at a pre-agreed price. Pricing by Amazon is perhaps based on a more mature model, as they have pioneered a part of this type of billing.
With security being the most used buzzword as the challenging part under cloud, can you elaborate the security framework defined around this model?
You are right. Security in cloud has been a point of discussion and rightly so. No one with a high stake on application and data will expose himself to additional risks by hosting them off-premise, where the control over resources may not be hundred percent. But, let us face it, security or lack of it is also a matter of perception. Existing solutions from major vendors are adequately secure. By deploying additional software for compliance and verifiable security responses, the cloud options can be effectively harnessed for most types of applications hosted in public and hybrid clouds. In any case, private cloud is almost as secure as any other existing system and has been found to be acceptable.
How will the procurement (license) model going to be impacted in this model?
If people move most of their applications to cloud, obviously, there will be lesser demand to buy local copies of the software for in-premise deployment. That means fewer copies of software licenses will be sold. But vendors offering products to be used in Software-as-Service mode may suitably adjust the pricing model to recover their cost and margin. So it is not easy to predict, how exactly the move to cloud will affect individual product streams. But the fact remains that the end-users will pay for the software usage in a fairer manner based on the actual quantum of use. Frills and unwanted features that are thrust upon users are forced to buy, in the existing scenario will be a thing of past.
Elaborate on the different business model evolved around cloud and which is the most popular?
As I stated little earlier, cloud computing is all about pay-per-use for all resources you may use in a very granular manner. There is a price attached to each unit of resource you may actually use/consume. Typically resources are classified into five categories such as CPU, memory, storage, program and network. Each one of these is metered accurately and contributes to the cost. There may also be a variable cost that service providers will charge based on measured Quality-of-Service (QOS) and Service-Level-Agreement (SLA). With so many dimensions in the pricing of services, the optimization strategies by the user and the provider will be complex. Let us wait and watch how it is going to evolve over a period of time. One can be sure that there will be many innovations here including many freebies.
Kindly furnish the names from key industry segments, who have deployed key applications on the cloud?
There are many established companies. Let us leave it to guess. It will be worthwhile to mention the popular everyday services we use from the likes of Google, Yahoo, MSN and Salesforce. They are all based on cloud technologies.