Advertisment

Big Blue. Grey Matter. White Noise

author-image
Preeti
New Update

NEW DELHI, INDIA:

Advertisment

Boy (in the movie ‘The Matrix): Do not try and bend the spoon. That's impossible. Instead... only try to realize the truth.

Neo asks: What truth?

Spoon boy replies: There is no spoon.

Advertisment

Neo: There is no spoon?

Boy answers: Then you'll see, that it is not the spoon that bends, it is only yourself.

“There is no spoon.” Interestingly the line echoes during a full-of-aporia-and-epiphany-in-the-same-breath conversation with Piyush Gupta,VP, Product Management, Information Management Division, IBM who assures you at the very onset that he is not just witty, cheerful and entertaining in his demeanour but also very very incisive when it comes to insights.

Advertisment

Ironically enough, he will tell you that he would not offer his customer a knife when only a spoon can suffice. Or for that matter, even a fork. His aversion to too much cutlery starkly stirs up a lot more curiosity in a world where ‘multi-pronged’ is de rigueur in every second enterprise pitch and beeps loudly on almost every new IT strategy dashboard. Widdershins?

What makes him stick to what some might hasten to paint as ‘plain vanilla’ and ‘boring’? Is it because ‘plain’ is also a synonym of ‘simple’ apart from being an antonym for ‘exciting’? Is it because we ask him how in-memory matters in a world prefixed with BI and suffixed with Big data and analytics? Is it because he feels that he would rather focus on ‘real needs’ than ‘perceived wants’ in today’s market? Is it because he reminds of the tag line ‘we are second but we try harder’ when he spares some attention to arch rivals stealing some pace in appliance-side innovation? Ahem, ahem, ‘innovation’ is another word he can make you have second thoughts about, by the way.

Is it that or simply his ardent view that ‘big’ should not equate to ‘complex’ but in fact, otherwise? Let’s find out why and whether he can make you pick your spoons to clink your glass in agreement here.

Advertisment

 

Since we would be talking at length on ‘Big Data’, can you first help clarify a moot point- With so much noise about ‘Big Data’ what makes it real and not just more hype?

Big Data is surely all over the stack, be it appliances, infrastructure or software. But IBM’s view is different from that of the industry. That’s incidentally where reality and hype part ways too. Industry’s view runs around Hadoop, for instance. But IBM’s view is to add value, ability and accelerator to Hadoop. Same way, the data warehousing piece still matters in the big data equation. The three Vs of Big Data: ‘Value, Velocity, Variety’ are equally vital in terms of real time data making real sense. Data warehousing takes care of existing data, Hadoop takes care of unstructured data while real-time analytics or analytics take it to another level.

Advertisment

Velocity. Let’s dwell on that for a moment. The industry is rife with excitement over in-memory breakthroughs which some dub as HANA and others as Exadata. Where and how does IBM PureSystems line strike a chord here?

Yes, appliances are an interesting realm. But this is not necessarily new stuff. We have always known how data processing graduates in speed depending on where it sits —from being the fastest inside memory to little slower in flash to different speeds in real memory on disk. Naturally if one can have all data fitted inside memory, everything would be faster. The concept is not new. Having said that, I would say SAP has done a good job marketing it well. We make hardware for HANA and it’s a good appliance actually. Objectively stating the downsides of in-memory, I would ask- what if your data exceeds the size of your memory? That’s why we have taken another posture, of going beyond memory. Here we allocate memory spaces to data depending on its temperature. To put it simply, the hottest bit of data goes inside memory, and the not so critical one can be in flash.

In-memory processing matters then?

Advertisment

The motivation is same for IBM as is for others. We came in with ‘Pure’ offerings which I agree is behind Oracle in terms of ‘who was early’, but the flip side is that we spent as much time in thinking it through and consequently we are better. What comes out when one compares Exadata in many ways to what you offer, specially when Oracle had an edge of time? They did put things ahead of time. But time taken to have either one operational can be contrasted easily. Between Exadata and our solutions, it swings between months to hours if you are talking the time to put the system in place and running. IBM’s is a much superior solution because we have waited and watched. Also, if you know what kind of workloads will fit, our solution will outweigh a box any day.

How is that different from Oracle is also offering?

We do not give too many options.

Advertisment

And that’s better? How?

I strongly believe that limiting choice is good in some cases. They are giving out ‘sporks’ (like a fork and spoon or may be a knife together). I would rather offer only a fork. What’s wrong with a pre-packaged box with a menu card to pick your options? Why should we inundate the customer table with soup spoons, dessert spoons, forks of all sizes, spoons of all variety and more when we clearly know he would not need them? Oracle’ s offering builds everything together in one utensil. I however would prefer giving my customer an instrument as per his need. If he is sipping soup, a spoon; and if he is eating spaghetti, a fork.

So choice is not good? Just some plain hot coffee as we sometime wish at giant coffee chains hoping and sighing if they would leave us alone?

Yes. Choice matters but not always. At times simplicity is all that a customer would be happy with. Give him what he needs and leave him at ease. When we put a new system out, we ensure everything runs smooth. That matters more. What is this whole deal about pre-integrated systems anyways? So much debate over it already. I will ask a question here. Why can’t we build some expertise into the system, like how much hardware or how much memory or how many clusters? Apriori information and parameters are given and made. It makes the post-deployment management really simple. One has to just call one console and solve everything. At the same time if you still are an enterprise that wants to make build something on its own, we can do that as well.

With analytics, and specially for customers who have invested a lot in existing pieces related to BI, or DW or other tools; ‘complexity’ can be daunting.

Making technology absorbable gets our attention. For some customers, the pick-choose-build approach works. For others, we can pre-build as per some use cases and that works really good for them. I would say it’s all our (the industry and other vendors’ and ours too) fault. It’s about the way we deliver technology. That has made maintenance, installation etc relatively complex. Why can’t he buy database from someone, an appliance from someone else and still have no headaches with a piecemeal approach? We will never leave a customer out, even if there’s a rival company’s system in place. We want everything to go smooth. Our surveys point out the irony that while CEOs are inclined towards doing new things, CIOs are busy with keeping the lights on. That says a lot.

Does it say that customers are still grappling with a lot of complexity?

Yes, and that common sense is something that can come without super-powers. A smart car manufacturer would tell you that while the driver’s seat gets more attention, marketing and work those other seats in foreign countries; the same cannot apply for India. Because in western countries, the customer usually is the one who drives the car most of the days of the week. In auto industry, even though a car is just the same machine, we see all types of cars — racing cars, off roaders, luxury sedans etc. The Model T days are over. Use cases are different. So that’s why when Oracle says- no matter what your question is, the answer is Exadata; IBM’s philosophy is fundamentally different. India’s market is different but ready. We are figuring out how to do things differently here, wherever required.

Has it been easy for IBM from a portfolio perspective? Galvanizing everything together?

This definitely called for a huge cultural shift as well. When different hardware, software etc divisions have to align together, rolled into one force, it entailed a change. To bring the ‘Pure’ line into market, there was a lot of co-operation that was needed across domains. CIOL Bureau

cio-insights