India has the opportunity to leapfrog: Rob Lee, CTO, Pure Storage

According to Rob Lee, Chief Technology Officer, Pure Storage, growing off a relatively smaller base, has an opportunity to take advantage of newer technologies and leapfrog, especially in the storage market.

DQ: Tell us about the Indian market and what kind of growth have you seen in the last few years. What about your future plans?

Rob Lee: Looking at the projections for the India storage market, it is much higher than enterprise storage overall. The numbers we have seen for the next five years' place the India forecast at a kind of mid-teens annual compound growth rate versus the storage market as a whole, somewhere in the mid-single digits. It is growing much faster, but still off a smaller base. 

This is one of the reasons why we entered the Indian market a year and a half to two years ago. Part of that is, firstly, as the Indian economy growth continues to drive, you would expect enterprise spending overall to follow suit. 

So, we think that is one of the big drivers for the forecast growth in the enterprise storage market. But then if you look at the enterprise storage market in most other parts of the world, still the vast majority of the storage spend is going to hard disk-based technologies versus Flash. We believe that India, growing off a relatively smaller base, has an opportunity to take advantage of newer technologies and leapfrog.

DQ: So, you think India is ready to leapfrog in this particular market?

Rob Lee: In terms of bypassing disk adoption into Flash and timing-wise, there are two things. The growth of the storage demand along with being able to offer Flash-based solutions at competitive costs to the cheapest disk system. If one can have all the benefits of Flash, 5th to 10th the space required, and power and cooling and energy, having it at the same cost as buying a disk up front, why wouldn’t they want to choose that?

DQ: If cost is not an issue, are there any other legacy issues?

Rob Lee: With our technology, we can neutralize cost as an issue. We couldn’t do that two years ago. Flash has to be way easier to implement. One, we have always prided ourselves on a very simple-to-use interface. We took a consumer-like mindset to enterprise technology. 

When you buy a phone, did you ever read the manual for your phone? You know how to swipe, you know how to pitch, and you learn the whole phone. We took that same mentality to designing our products. And if you talk to our customers, they'll tell you that we win deals-POCs because we can install in 15 minutes, and the customer gives up with a competitor after trying for two weeks.

But where we set ourselves apart with our technology versus disk from an operations point of view, is reliability. Our systems and our hardware can be 10-30 times more reliable than disk-based systems. It’s because of our Flash technology and our density, we typically ship about a fifth less of the equipment to serve the same storage need. 

DQ: Coming to specific products like Pure Storage and Portworx data services, products, and areas, how has India taken to it any differently? What kind of products are in a large growth area?

Rob Lee: Especially in the commercial SMB, the Indian market is some parts or more price sensitive. We expect some of our more cost-optimized offerings, our FlashArray C, our FlashBlade and FlashArray E, and our disk takeout offerings, to be of higher demand, especially as we look at the commercial and SMB, probably the smaller capacity type systems. There is a lot of potential for the Portworx suite of products. Looking at a lot of the high-tech sector in India, leapfrogging technology, and new software that is being built, starting with containers, presents a great opportunity for something like Portworx, even though it is still maybe early days.

DQ: How do enterprises and SMBs differ, and what is your SMB focus? What is your go-to market strategy for SMBs in India?

Rob Lee: Our company started with a very heavy commercial and SMB focus. That is partly because storage has been an industry that has been very slow to change. SMBs and commercials tend to be easier and quicker to adopt newer things. We have a very strong base there. That is where a lot of our products are very well suited. 

When we look at the enterprise and what the enterprise needs, they tend to be things like: How can I automate better? How can I manage at scale better? It's great that you made one thing easy to manage, but if we must do something easy and simple 100 times, it is not that easy and simple anymore. But again, that is where the evolution of our enterprise maturity and usability has been something we have been growing over the last five years. 

From a go-to market strategy perspective, the key difference is in where you target your direct sales efforts and where you leverage the channel and the partner community. What we have done in the US and other geographies quite successfully is put a lot of focus on enabling our partner community to serve a lot more of the SMB and commercial market and focus our direct touch sellers on the enterprise sales.

DQ: What is your cloud strategy? What about CSP partners? What are the kinds of partnerships you are looking for in India?

Rob Lee: Looking at cloud partnerships and cloud strategy, there are a couple of different things in there. The primary ways that we help customers run better on the public clouds are either our Cloud Block Store product or the Portworx product set. They both provide a customer with the means to have the same storage experience across different environments, whether it is on prem, or in Azure, AWS, or other clouds. Where they differ is Cloud Block Store is more focused on traditional sand-type environments, more traditional applications, VMware, and SAP, whereas Portworx is more focused on cloud-native container-based applications. But both provide the same kind of portability. We have had great success on both fronts. 

DQ: India has a lot of R&D and innovation centres. Are you looking at that aspect in terms of driving R&D from India?

Rob Lee: We as a company have historically been a very Bay Area-focused engineering organization. But a couple of years ago, we very consciously started expanding our global site strategy. We first started by building an R&D centre in Prague, and then Bangalore is our third site. And our strategy in how we are running global R&D is a bit different than the rest of the tech industry. A lot of companies will open satellite sites, but they will maintain more direction, control, and design decisions centrally. We view our global site strategy as very different. We want to create independent ownership of large-scale functionality, parts of the product, and potentially even entire products in our global sites. We want it to be more of a network of R&D sites, as opposed to Hub and Spoke. 

DQ: What are global and Indian tech trends? 

Rob Lee: From a global point of view, there is an explosion going on in terms of data growth and a change in how data is being produced and consumed. Looking at how data gets produced, it is scaled with GDP. You go to Amazon, you buy something, there is a credit card transaction, and the database grows. You hire an employee, you provision a VM, and it grows. 

Look at how data is being generated today. There are only 8 billion people, but there are 12 billion phones. How much data is being generated, it is all being generated by devices. The scaling factors are quite different. But then you combine that with newer technologies. Now along with a means to generate an incredible amount of data, but there is also a demand to go and consume it. Those are clear trends. Those have implications in terms of software needs, computer needs, and certainly data storage needs. That is the biggest set of macro trends. 

We think that that is very constructive for Pure, mostly because if you look at that incredible amount of data 20 years ago, there were places that were storing huge sums of data, but they put it on tape or they put it off in the corner because you never had to use it. The only optimization function was to reduce cost. Well now you have to actually make use of it. It has got to be warm. That's where our technology, our products and services, and the fact that we can do that very efficiently in terms of space and power is very compelling. 

These trends are also exacerbating some limitations that are out there. It is very rare that I talk to customers these days that are not under some sort of pressure from space and power and expansion. I was in Singapore yesterday the last couple of days meeting with clients, and Singapore has a moratorium on data centre construction. The government said you are not allowed to create more data centres due to problems of space and power. And they are not alone. Some of it may be climate and carbon-based, but it is more production. We just do not have the energy. 

For Pure specific to the Indian market, we would expect India to be a little bit on the leading edge of some of the newer technologies. Part of that has to do with India having an incredible labour base of high-tech workers and an educational system and expertise here. India has a lot of foreign investment. It also has a lot of policy initiatives to drive in that direction. It has all the fundamental ingredients. But India also has less of a legacy base to leave behind. Everybody else, there is some amount of inventor’s dilemma or innovators dilemma.

DQ: Because we have talked about data explosion, and the hunger for more structured data analytics, where does ChatGPT or generative AI feature in all of this?

Rob Lee: I would differentiate ChatGPT from generative AI. I believe ChatGPT with its large language models is going to change how everybody operates. We are using the technology internally just to optimize some of our functions. But from a data perspective, it is not actually that data heavy. If you look at text data and the entire corpus of the text data that was put into ChatGPT, it was like a couple of hundred terabytes. One of our small arrays could hold that. But when you dig into the infrastructure and the internals of how many GPUs are required to go train that or the intermediate data steps, that is where we can help a lot. But the underlying data set is not that large. On the other hand, generative AI as a technique is going to be applied to all kinds of other data sets.