The Complete Magazine on Open Source

“We get a lot of good talent when it comes to open source”

SHARE
/ 260 0

Piyush Jha

Ask a CTO about what the biggest problem is when it comes to using and implementing open source, and the answer would be ‘manpower crunch’. But Piyush Jha, AVP, product engineering, GlobalLogic, has a different opinion. He finds enough talent to carry out his open source projects. In fact, he thinks that the modern day techies vouch for open source and the power they get with it. Diksha P. Gupta from Open Source For You spoke to Jha about his open source implementations, while inviting him to Open Source India 2015. Excerpts:

Q Do you use open source technologies at GlobalLogic?
A lot! We develop a lot of components over open source. My background has been mostly in Java. When we try and decide on the technology for a project, we often go with open source since that’s a pretty comfortable zone for us. Apart from Java, we also work with open source languages like Ruby, Python, etc.

Q When you start developing a product, what’s your first choice?
We work with a lot of partners and develop products for them. Being product developers, we need to get into the ecosystem of our partners, and try and see what they are doing. So, we generally try and propose a tech stack that mostly matches with the tech stack of the partner. Let’s say a company is a complete .NET enterprise, then we have to propose technologies that are closer to .NET. We try and propose a tech stack that matches theirs. But if the decision is really open, then we carry out research on which technology suits their needs better.

Q When it comes to your clients, how much of acceptance is there for open source?
There is a lot of acceptance. What the clients want is that the service provider or product development partner should be well versed in the particular technology that’s being recommended and be able to back that up fully. So, if we show them that we do have enough expertise in what we are talking about, then things get easier. Open source won’t get the desired boost if we do not have the backing of enterprise-led activation. So, if we have some successful case studies with respect to open source, then convincing clients is not a challenge.

Q How has your journey with open source been so far?
It’s been extremely good. Our engineers are pretty comfortable in approaching forums, looking for a solution, trying to resolve a problem that has not been resolved before and getting their problems solved. People go for closed source solutions or tried-and-tested solutions if they are not confident about finding a way out of a problem. Our team is pretty good at going to forums to seek solutions, and it tries to find out how to do what others have not done before.

Q What are the pre-requisites that one should have before going for an open source solution?
In general, you have to break the problem into parts. So, even if the specific product that you are attempting to implement is something that you haven’t done before, you would like to try and see whether its components or pieces have been already worked on before or not. If the pieces have already been worked on, it is a cakewalk. But if they haven’t been worked on, then you would pick up a couple of pieces that you find slightly challenging or those that slightly deviate from normal coding and try and do a small proof-of-concept (PoC) around it before adapting it as the solution for the complete team or the clients to follow. So, these PoCs are pretty relevant in the sense that they take away a lot of the risks.
As an example, we are currently working with a leading retailer in India to try and automate its warehouse management system. At present, its systems are mostly on paper or Excel spreadsheet – that is how the picking, packing, shipping process is carried out. We are in talks with the firm to try and automate the process. When we proposed the solution, we found a couple of elements that they are not really confident about and which they haven’t worked on before. And those are the parts of the solution that were pretty exciting to them. We took a step back and told them that these are parts in the complete solution that we haven’t really done before so let us get a proof-of-concept on these points. So, we are currently doing a one-month project rather than a complete one. Once we are done and agreed on what the thresholds and boundaries would be, we will go for the complete solution.

Q What are the advantages of using open source technology?
Cost is one definite advantage. Apart from that, we see that generally, updates come pretty fast in open source as compared to a company that would be developing a particular product. There’s an army of people who support it and keep contributing to it, which makes it pretty easy for any problem or any new idea that comes in to get implemented in the open source project. So, updates, adaptations or innovations come out first in the open source world. Whatever is coming in as a standard solution gets adapted pretty fast but besides that, there are quite a few innovations that come out and we get the first-mover advantage when we roll out innovations that people haven’t seen.

Q According to you, what are the disadvantages of using open source technology?
You don’t have solutions readily available all the time. At times, the support is limited. And you could get into some tight corners which are difficult to come out of. So, it’s not easy to get a vendor from that corner for support. It would be slightly complicated since you have to work your way into forums, try and get your queries answered, etc. Nobody might have worked on the solution you have in mind so you might need to work around the problem.

Q Do you get enough talent to work on open source based projects?
The open source projects are getting increasingly exciting. The techies and even those who’ve just passed out of college are excited about open source and are willing to try their hands at it. Moreover, it’s not really tough to motivate people to work with open source.

Q Are these people talented and skilled enough?
Yes. Because the geeks are more excited about working on open source, we get a lot of good talent when getting started on open source projects. Interestingly, they keep themselves updated on what is happening in the world and would know a lot of stuff that our clients may not have heard of. So, it’s easy to get people. Also, I believe that the number of people in open source is either equal to or higher than those using proprietary technologies.

Q You are one of the few CTOs to say so. Most people feel that talent is one of their biggest problems…
We are seeing a lot of traction in the Java world today. There are a lot of firms that are working more on Java currently. If we look at Indian e-commerce players, all of them are Java sourced, like Flipkart, Snapdeal or Myntra and if you look at their tech stacks, it is a lot of open source. You might think I am one of the few telling that talent is easy to find, but we do have quite a few companies who are actually able to find and use this talent.

Q When did your tryst with open source technology begin?
We started to do small projects during my third year of college in IIT. We did a lot of niche PoCs and then I did my internship at the University of Paris. There were a couple of professors there who were extremely pro-Java and pro-open source. So, working with them for three months actually exposed me to a lot that I was unaware of earlier. That experience got me more interested. Over the first four years of my job, I worked on five or six different languages including C++, Java, Perl and Ruby.

Q What’s your primary language right now?
None. I am supposed to be a pro, across the board. To me, there are a lot of solutions that themselves shout out for certain languages to be implemented. So, if you want to build an extremely scalable enterprise solution that supports 10,000 concurrent users, or a million users or hits at a time, per day, you would probably go with Java. However, if you want to go to market pretty quickly with a website, and your investment pressures or your go-to-market strategy forces you to get that done within 1.5 months, you will probably not go with Java and select Ruby instead. So, it really depends on what your product vision is and what’s the technology that your product is looking for. The decision is taken on that basis. Basically, the technology choice, more often than not, lies in the problem statement itself.

Q How easy or difficult is it for you to convince the management about using open source?
We are 50 per cent open source and 50 per cent proprietary. So, that has never been an issue. The management has always been convinced about it. Good techies guys have founded this company and have been around for quite some time, so open source has always been in the company’s blood. Even today, our management is pretty supportive of what the technology team says. When members of my engineering team back up the solution and say that they can do it, we rarely face any resistance from the management.

Q Are there any community contributions that your people do?
Yes. A lot of our people are pretty active on blogs and forums. At times, the CTO office is involved in global capacity contribution but at an individual level too, we contribute pretty often. In GlobalLogic, senior people have a part of their salary as a bonus component. More often than not, 15-20 per cent of that component is earned by their contribution to the blogging world, the open source world and the world of technology. So, we have a concept wherein a part of the bonus is reserved for performances or contributions that they have made to the project, to the company or to the technology world, in general.

Q Twelve years back when you started, was open source technology as good as it is today? What were the challenges then?
I think it was even better than today. Now, there are a lot of proprietary solutions available and the world has evolved a lot. Even Java solutions today are proprietary. At that time, Sun was completely backed by open source and, today, I would say that the market is 25-30 per cent open source. At that time, it was at least 50 per cent open source. There was a lull that we saw in the middle. But in the past two years, with the mobile revolution coming in, the open source wave has started to pick up momentum again.

Q How much do you see it changing in another 3-5 years?
I think it’s in a stable state. It is here to stay because the IT world today is completely oriented around innovation. There’s a lot of competition around products. Let’s take the example of e-commerce products. There are at least 10 portals in India, of which three or four are in the top league and they are pretty close in terms of their competitiveness. Their only differentiation is in the kind of features and the ease of use they provide.
Any company amongst these that comes up with two or three attractive features first, gets the advantage of taking a leap forward. And in such sectors, whenever you mention innovation, open source is kind of a front-runner, by default, . There are a lot of people and good thinkers around who think up something, write code for it and put it in. So, whenever you talk about innovation, a lot that’s new and path-breaking would be open source. So, I think in the next two-three years, we will see a lot of evolution in open source.

Q What are the areas in which you see open source growing a lot?
If you talk about domains, retail e-commerce, mobility and media will continue to have a lot of open source elements. IoT is an extremely vast playground, where open source will play a very large role because that’s a field which is still evolving. There will be a lot of media solutions. Media is changing a lot. There are very few people who read news on a desktop today. It’s all mobile. People read the news, like, share, cut elements, etc, all on-the-go; so this is another area that will have a lot of open source contribution.

Q Is open source still a back-end property?
I really don’t see any reason why that would be the case. In fact, many of the front-end operations use mostly open source, because it’s much more feasible to do so.