Cloud-Based Apps To Constitute Significant Chunk Of IT Spend

Sharad Sharma, while in the capacity as the CEO of Yahoo R&D India, in conversation with Biztech2.0 shared his insights on the cloud-based model.

What are some of the latest developments at Yahoo R&D?

A lot is happening at Yahoo R&D, especially in India. We are doing a number of things on the multimedia search front. Last year we launched a product called ‘Glue Search,’ first in India, which was later introduced in the US. A lot of our work is focused on the advertising spend side, particularly on search monetisation and a fair amount of work is happening with our consumer centric properties. Coming to the cloud there is a considerable amount of work which being done in India particularly with regard to Hadoop, which is a distributed file system. So, Yahoo’s approach to cloud is really at the infrastructural level, for consumption within Yahoo and not for external consumption at this point of time.

What is Hadoop?

As mentioned earlier, Hadoop is a distributed file system. When you have an application that can be processed in parallel and is data incentive, then a system such as Hadoop would be easy to use and the apt choice. Hadoop is Java enabled and yet it runs on the cloud or the grid. Yahoo has taken the lead in open source development of Hadoop, which has been very successful. Academics and universities have provided innovative expertise and participation in this development in addition to what Yahoo is doing.

What is your take on cloud computing?

To my mind it is as big a shift as was the shift of Mainframes to client-server computing. It is therefore a threat as well as an opportunity. Just as some of the mainframe application companies did not succeed in the client server world, it is quite possible that some of the traditional companies may not be successful in the cloud world. I am excited about this model as it has the potential to unlock the consumption of IT within the SMB sector in India. All of us tend to think it will happen very quickly, but I think it is going to take close to four to five years for it to play out. I think the direction is set and from this point onward things are only going to gradually move to a world where cloud and cloud-based applications will become a larger percentage of the IT spend.

What kind of applications do you think will move to the cloud?

There are two ways to look at the cloud, one is the horizontal cloud, which many people now call as platform as a service, while the other constitutes applications that are built on top of the cloud. Employees within the company, as well as consumers outside the enterprise can use the latter. I think the range of applications is fairly broad, including communication applications such as webex and web-mail to the productivity application that Google is driving. These range of applications also include the likes of Salesforce automation, low-end ERP solutions, financial analysis applications to more high-end applications such as BI and decision support.

What are some of the major apprehensions in the adoption of cloud computing?

The traditional apprehensions in the adoption of cloud computing are related to availability and security. For example, situations such as the networks and servers could put the enterprise in jeopardy. In India, often there are concerns about network availability and not just the data centre availability. Due to credit card and other online frauds people have growing concerns about loss of control over data in a cloud model. However, these kind of concerns are getting resolved gradually.

What kind of role do you think will open source play in a cloud model?

Since we are in the early stages of evolution of the market right now, I think there are two possibilities for the cloud model, one is that there could be proprietary development, which is platform as service; examples of this would be Force.com and Google’s app exchange. An alternative environment that could develop will be based more on open source, which would include participation of people with the addition of layers of capabilities in an open source environment. If the latter was to happen the there are chances that Yahoo driven Hadoop could be taken with some other open source implementation for application development. Thus, this may end up in the building of a new stack, which is built on open source technologies. I think it will take two to three years for concrete and value-based models to emerge but both are viable options at this time.

Do you see cloud posing a threat to traditional IT vendors?

I think clearly the application developers will love cloud computing because this in a way disintermediates the large system integrators from the chain and the end-users will also favour it due to its numerous benefits. Because of the disintermediation impact, system integrators could be at risk and therefore have to start thinking about how they can reorganise their business model to leverage this model as they go forward. It isn’t as though they don’t have a future, but the future is going to be very different and they will have to move up the value chain, offer more consulting services and productising their service packages using cloud computing.

What opportunities do you see for Indian IT vendors in the cloud space?

Traditional companies who have benefited from the old model are trying to embrace cloud computing and I think some of them will succeed. However there is enough and more space for new companies to come in. This is something that is really a huge opportunity for Indian companies. Indian companies have already started making their mark in this space and a good example of this is Zoho. This is bringing companies from across the world together and I think Indian companies have the potential to do exceedingly well.