Saturday, December 21, 2013

Delhi: Memories, Objects and Change

Dec 22, 2013



Delhi, India.  I shouldn't say Delhi.  I should say Dilli.  I found myself here with a few free hours today, out of an *incredibly* hectic week.  So I went by some old spots, Bengali market, Hanuman Mandir, Connaught Place, Lajpat Nagar, ....  The brain is immediately drawn to the sights, so different now, yet still familiar.  The colors, the haze, the crowds, the structures, the spaces.  And beyond the sights, the sounds, the tastes, and, especially, the smells.  The smoky, dusty, musky air.  The fragrances of flowers being sold and foods being cooked, and the foul smells of garbage.  All mixed into an unforgettable reminder of the ephemeral present, that is also, yet, timeless.  More than any other sense, the smell takes you back.  But back where?  I remember being here 30 years ago, shortly after Delhi had seen a great renewal, in preparation of the Asian games in 1982.  The structures are still there, but they are different.  There is Talkatora stadium.  There is RML hospital.  Wow, Bangla Sahib is so different now.  My mom used to go to these places.  This is her Dilli.  And of the rest of us.  I guess most of it is from memory, perhaps the rest is my imagination?  But it isn't only the structures that are different.  It is also us.  We change.  Indeed when we think about it, change is all there is.  The constancy of the twirl, the great movement, both human and beyond, that I see around me, was in that sense also there 30 years ago, just as it is today, and yet it is different.  What is it?  Is it an activity?  Is it nature continually transforming the objects around us in a kind of eternal dance?

The more I think about it, the more convinced I am that while we think of the world in terms of objects, things, entities, etc. indeed all these are temporary constructs.  Activities, and their change, seems to be all that is going on around us.  And everything that comes across as objects, I believe, is simply a set of activities in progress that our senses "snapshot" into an object, temporarily, ephemerally, persisting/materializing it in our memories, as though it is a fixed, permanent thing.  Even though it isn't.  Our societies, even our languages, seem to be geared towards "things", not "activities", geared towards particles and objects, not waves and processes.  THe plate of chili chow mein in front of me, seems to be an object, but in fact is something that was flour and water and a bunch of other things that came together into an activity for a short period of time and then disappeared.  It is just a temporary materialization of something we observe and experience.  SOmething that exists only for a moment, and in our senses.

I get the distinct sense that we must improve our ability to articulate actions, activities, processes, and think in these terms more so than in terms of objects, and to think of objects as transient materializations of activities.  And as I think of this, my thoughts drift off to computing.  Us computer scientists, and IT practitioners, are horrible, and horribly primitive, at articulating actions, activities, processes.  Even in purer object-oriented languages, most of the code seems to be about articulating actions.  Whether it is software actions, like "book me a flight to London" or "balance my checkbook" or "repair a customer's credit" or whatever, or more "concrete" actions like commands to a robot to "go to Vishal's office with a cup of tea".  Our ability to articulate actions succinctly and precisely, being able to extend, compose, project on or decompose actions is extremely primitive.  We are still in the dark ages in this regard, and we must improve.  This is one of my endeavors with the work on River and much more needs to be done here.

But all that is for a different day.  Today is about Dilli.  And to head out with some friends on a last evening here for this trip, to observe, and participate in, some activities, to make some memories, in a memorable place.  For Delhi is more than memorable, it is a permanent memory.

-- Vishal

Sunday, November 3, 2013

Seeing the Lights: Some Frontiers for Enterprises


It is the morning of Diwali, and already Diwali evening in my Matrabhoomi (country of birth).  As Indians around the world look ahead to lighting Diyas in their houses, and lighting up the skies with fireworks, I've been missing being back home, missing the sights, the sounds and the tastes, and most importantly the lights.  The colors, the glow, the patterns, the beauty, the magic.  Good news is, here in our adopted homeland, there is plenty of Indian diaspora, and there will be, I'm sure, much fun and fireworks for all of us Indians, and the Indians at heart.

But Diwali is about more than the lights outside, and indeed much more about the lights within.  To see, and to reflect on, what is important.  How we should shape and focus our efforts, our intents and goals, and our time.  To work on that which matters most.  The "good", and not the "evil".  I've been thinking about these matters lately.  Thinking that the leading institutions of our time don't do enough to move the ball fwd for humanity, focusing rather on advancing their own agendas incrementally, or worse, their egos and reputations instead.  The big consumer web companies seem driven by taking over more and more of our privacy, and people seem to have a general lack of understanding what they are giving up to receive these "free" services, and end up sharing their information trail to be monetized by others for all it is worth.  Leading companies in all major industries seem focused on dealing with the new technological realities that have enabled new economic models that are profoundly disruptive to the ways they've done business for decades or even longer.  And often an early casualty of this is  investments in research and fundamental work in advancing the important areas.  At a time when we know so little of our world, so little of our own self, our brains, diseases, energy, etc., even less is being done to invest in unconstrained/unburdened and long-term investigations into these important matters.  Learning would be the missing link, but almost all work on rethinking education seems to also be about commercial endeavors that focus on new "learning platforms", rather than on improving our ability to learn.

When I think about the work that can be done, e.g. in the software industry, several things come to mind:

1. Learning of course.  Better environments for learning, a culture of education, so businesses, and employees can better adapt to, and respond to changing circumstances.  This has to be one of the big endeavors in all businesses.  I am working to make this happen at SAP.  Both in our own company's culture, as well as learning related products that SAP can bring to businesses.  And beyond learning, in better ways to communicate, to share, to understand, and to design.  All are needed.

2. Economics.  There is a lot to be done in better understanding economics of the key issues of our time, and to transform our value-generation towards long-term sustainable models.  I believe there are ome key economic tradeoffs that it would behoove us to understand more deeply: e.g. the economics owning assets vs sharing/renting these.  The so-called "rental economy" or "collaborative consumption".  A better understanding of the thresholds at which it makes sense to own something vs to rent/share it, would be very beneficial.  There are similar thresholds in (i) make to order vs make to stock, both in manufacturing as well as in other industries, in (ii) reconstructing things vs simply renovating these, as well as in (iii) when to use a central shared service vs decentralizing autonomous teams.  I've been exploring these issues, and also sharing some of this work in a Stanford Computer Science Class that I'm helping one of my PhD advisors with.  I believe that there is a kind of threshold (a "Sikka Threshold" if you will :-)) that governs all of these tradeoffs, and they all have a structural similarity to each other.

3. Software. And of course, there is tons of room in improving the software situation around us, especially as software becomes the vehicle to bring about, to instrument, this latest great transformation unfolding around us.  Improvements are needed in many aspects of software

a. Such as the nature of software platforms and systems.  At SAP, we've been busy building on HANA's early success to further develop new architectures and landscape designs for enterprises.  We are still in early days of understanding elastic, intelligent, scalable systems.  For instance, the management of time in systems is still quite primitive, typical ACID behavior in databases forces distributed systems to go thru tremendous contortions, when even quantum physics has taught us that absolute time across distances and rigid notions of consistency are not only futile, they are not necessary.  I am writing a paper about relative consistency, and much more simplified notions of time and management of data reliably in distributed, scalable, systems.

b. Such as rethinking the software experience itself.  The software developers' experience leaves a *lot* to be desired and there are fundamental ways in which it can be improved.  We've been working on some of these.  From ways to make software development real-time and totally interactive/responsive, to enabling end-users to do more of the programming work themselves, from better understanding, and articulating, actions, to better ways to articulate software's abilities so strangers (alan would say "alients") can read and understand our code.  In particular, the notion of activities is one of these areas.  We understand so little about how to better articulate actions and processes in richer, yet efficient processes. (I am also doing some work on this area,  there hasn't been a significant advance on articulating activities richly and efficiently, and expect to write a paper on this to a conference/publication this winter).

c. Such as new, unprecedented industry applications.  Our notions of enterprise applications have become significantly limited over the decades.  It is time for a rethink.  As almost all major industries get disrupted or totally transformed by technology, it is time to build the great, intelligent, adaptive, applications, in unprecedented new areas of business.  But especially the purposeful ones.  The ones that move us forward, as they help bring to life the power of software to transform our world and our lives.

I think businesses, corporations, have a tremendous ability to influence the world, to shape the world of the future, and it must be our endeavor to do more fundamental work, more of the important things, the purposeful things, especially ones whose value isn't obvious in the 90-day rhythms and the lenses of the traditional financial metrics that seems to consume most large companies.

So as we get ready to light up the Diyas, and light up the skies with fireworks, I think it is even more important, and worthwhile, to reflect on, and to pursue that other light, the one that's inside of us.  Towards the end of his great book, Siddhartha, Hermann Hesse's title character says to his lifelong friend Govinda, both now old men, that knowledge can be taught, but not wisdom.  That each one of us must find that for ourselves.  Here's to wishing at this Diwali, that each one of us finds, and pursues, and achieves, that light within, our own, unique, wisdom...

Vishal


Monday, October 21, 2013

SAP, Software, and Amplifying Human Potential: Some Thoughts on the eve of TechEd


4 years ago today, I learnt of the death of my dear friend Ranjan Das.  He passed away unexpectedly, far too young.  As I was looking back on the last 4 years, and looking ahead to our TechEd conference that begins in Las Vegas today, I was stunned to realize the obvious: how time flies by so quickly, and how the things we cherish are the things that are timeless, in that these are the activities that last the test of time, and also the activities that so engross us that one loses track of time in.  Both happen to not be about the superficial things in life, but about things that touch us and that matter to us in a deeply personal way.

Around this time 4 years ago I started the HANA project.  We had worked on it for years already (I'd started the in-memory db work in 2002, and came up with the name HANA in 2006), but it wasn't until the fall of 2009 that we finally convinced SAP's management that HANA needed to be built and the time was now, and we started the HANA development project in October.  The day I learnt of Ranjan's death, I was in Walldorf, and this was the day Franz, the core HANA leadership and I sat together and decided that HANA would run under existing applications, both ByDesign and also the Suite, in addition to serving analytics, and all kinds of new applications.  That it would carry the load of the new and the old.  Both, simultaneously.  I felt that this was our burden to help renew SAP, a challenge Hasso had laid out to me earlier in the year.  I remember walking to my flight that evening, at Frankfurt airport, telling Hasso on the phone that this is what we decided, and he was very happy about it.

Ever since those early, heady, days, it has been a hell of a journey.  We built HANA in record time, and released it to customers on Dec 1, 2010.  She went GA on June 20, 2011.  Ever since she's been nothing short of a revolution.  We recently crossed a billion dollars in HANA revenue, by far, by a wide wide margin the fastest growing product we've ever seen in SAP, quite likely also in the industry.  This success is a result of HANA's technological capabilities, and the breakthrough benefits these result in.  Much has been written about this, by me and many others (saphana.com has plenty of background).  HANA, at its heart, represents a rethinking of the relational database, a reinvention of it, to reflect both,
(a) the new hardware reality of super-affordable x86 based machines that combine very powerful multi-core processors with the super-fast access to data in large memories that are now available in DRAM, and
(b) the new ideas in in-memory structures, especially the column store, the newly designed highly parallel structures and operators, and tons of new ideas in database technology.
And this combination enables us to bring value to the enterprise in totally new ways.  I wrote a paper at the ICDE conference this year, articulating a new way to represent this value, in scenarios that bring together data volume, speed and complexity in unprecedented ways.  HANA's value also enables us to rethink the application reality in the layer above HANA, both by simplifying and accelerating the existing applications, by refactoring these, as well as building totally new applications.  Many that were often not possible before.

And so all the database world seems to have woken up to this new reality.  IBM's Blu, MSFT's recently announced Hekaton and Oracle's recently announced 12c in-memory among others.  And yet when I look at the public material on these, it seems clear that they could have done much better, they could have done so much more.  The main point of HANA is a single columnar store, where the transactions go into memory, and are available instantly and as-is for analytics, even deep complex questions, due to the power of massive parallelism, and this point seems to have been fundamentally missed by these approaches.

But while the competition misses the point, we have taken the ball fwd, and taken major steps to make HANA our platform for enterprise applications.  We have added all kinds of interesting capabilities into HANA, from various application libraries for statistics, planning and business functions, to middleware and integration capabilities, to now even a complete set of application serving capabilities, so we can run and deliver entire applications directly from HANA (and, of course, if a developer wants to build apps in their favorite platform and simply integrate with HANA, they have total freedom in doing so).  So on the basis of these capabilities, we have been moving every single SAP product to HANA.  From the Business SUite to B1, from ByDesign to Business Objects, from Success Factors and Ariba to Sales OnDemand and Hybris.  Every one.

But as we look at this platform, and the capabilities that it affords us, and take the broader view, there is a sense that it must be about more.  As Alan Kay always reminded me, the future cannot only be an increment of the past.  If all we did with this platform, was renew things we already knew, we'd have fundamentally missed the point.  It must be about more.  About enabling new capabilities, building new things, great new apps that help transform the world with real-time software.  Ones that are purposeful, ones that empower us and inspire us.

The last time Ranjan and I were together, we'd spent 6 hours in a flight from orlando to SFO, during which he kept bugging me; he must have asked me a 1000 times about what the next big thing was, and that conversation led us to the point where I told him I think it was the ability to build the truly next-generation amazing applications around design, and creation, enabling the acts of creativity by companies, to help them truly find their purpose.  And we'd talked about many companies in India (in particular Mahindra and Mahindra -- Anand had been a very big influence in Ranjan's life) and how they would benefit from going beyond outsourcing and consulting, towards truly creating innovation.

Today we are doing lots of these.  Abdul and his teams, Thomas Torf, Priya, Alan Southall, Ritika, Prasad and others, work with customers to find truly important, meaningful, purposeful areas for them, that need to be transformed with software, and help bring these to life.  From end-user clientelling and responsive supply chains for Burberry, to real-time signal detection for EBay's analysts, from predictive maintenance for John Deere's machines, to forecasting and optimizations for NongFu Spring, Mitsui and other companies in Asia, from better oil and energy exploration, to personalized management of energy by billions of consumers.  And our team in India, led by Gansu and his gang, are working with individual milk providers to see their revenues from their milk-production directly for the first time, without corrupt middle-layers in the system, thanks to the power of  their Aadhar identities and HANA.  And beyond SAP, today we crossed a great milestone.  More than 1000 startup companies are now building their products on HANA.  More than 35 already have products in the market, addressing all kinds of needs, in all kinds of areas.  It is an unbelievable example of bringing the power of technology to help enable the empowerment of end-users.  Great technologies, from the bicycle to Gutenberg's printer, from the surfboard to the internet, have always aspired to, and managed to achieve, the amplification of the human ability, our intellect, our senses, our purpose.

At this TechEd, I am hopeful that we can share with our ecosystem our roadmap and direction, not only for our Task A, our technology and platform, but about how it can help us find our greater purpose, help us build the great applications that empower us, that amplify our reach, and enable us to do more, and even more importantly, help us to continually learn, and adapt and evolve, as we continually get better at attaining our intents, our purposes.  Ranjan would have been really proud...

-- Vishal

Sunday, January 20, 2013

January 19, 2013

I write this at the end of a difficult week, one full of sadness and reflection.  It started on a high, amid 3250 sales leaders at our FKOM in Singapore.  But was quickly followed by the devastating news of the death of two very special colleagues: Andreas Raab and Dean Jacobs.  I have never written an obituary before, and this isn't one.  Rather it is my attempt to piece together what their lives, and now their passing, have meant, and what I've tried to learn from this, after a few fragmented moments of reflection.

Andreas was a key developer of Squeak, under Alan Kay's guidance, and a distinguished member of our technical team at SAP.  He died suddenly, and abruptly, earlier this week.  And Dean, or Deano as I called him for years, another distinguished colleague and dear friend, succumbed to cancer, but not before giving it a hell of a fight.

All week, since absorbing these two hits, I've found myself wandering along, somewhat numb, wondering just how fragile, and fleeting, transient, life is.  How quickly, and abruptly, and certainly, it ends.  And how we are never prepared, even when we know.  How complacent we are, assuming that there is a tomorrow, and carrying on with minutiae and trivia, knowing, certainly in the back of our heads, if not in the front, that these amount to nothing.  How much energy we waste chasing after ghosts, fighting off stupidity, even when we are better off ignoring it, how much time we spend mired in nonsense, being slowed down by the viscosity of the inane and the mundane.  Unaware that moments of joy, and togetherness, and love, and passion, and giving, and creating, and being in touch with the nature within, and the nature without, constitute  precisely the intransience, and the permanence, that we seek, and yet assume for granted and ignore for the shallow and the meaningless.  And yet every once in a while, even if far too rarely, these truths shine through in our work.

After 9/11, Andreas wrote in his blog:



Dear friends and collegues,

The shocking incidents of today make it important for me to say two things: First of all, I wish to express my sadness about what happened in New York and D.C. and I am sure that all of the World is with the U.S. in this hour.

The second issue, which is actually far more important, is that we are in fact working here for a better future - a future in which such horrible incidents don't happen, a future in which our children will live and learn in peace. Computers - the internet - can help to understand other cultures better, can help to understand problems of regions far away better, can help to raise our attention to both, tragedies and threats from parts of this world seemingly far away.

Let us not get distracted by these horrible incidents. Let us work for a better future for all of us, and our children.

My prayers are with the families of all the people in the New York and D.C. area. Although the world will never be as it was yesterday, we can still work to make it a *better* world than before. What I've seen and heard today is in fact giving more hope than one would expect in such an hour.

- Andreas


Andreas, you were right.  Computing technologies are still early, with a promise to improve our lot that is far and wide.  We can, and must, continue to work for a better future for all of us.

Deano once patiently heard me out on an idea I had, back in 2008, and immediately called it VINA.  Those of us who know why I'd named my product HANA, can surmise what VINA would be an acronym of (these were different things).  He told me to pursue it with all vigor and passion I could, and even wrote up a two page plan and description for it.  He almost single-handedly woke SAP up to some harsh realities of the Cloud world.  His observation, that about 2000 1TB DRAM servers could hold all the energy consumption data and compute power to enable more than a billion people around the world (customers of SAP's utilities customers) to play with it, and take better control of their energy destiny.

Both voices now silent, taken far too soon, their dreams far from finished.  And yet both lived lives of passion, and love, and creativity and curiosity.

Jiddu Krishnamurti once said, when reflecting on death, that one cannot fully understand death without understanding life, and that "One may try to give meaning to life, as most people do, saying life is this, or life must be that, but putting aside all these romantic, illusory, idealistic nonsenses, life is one's daily sorrow, its competition, despair, depression, agony - with the occasional flash of beauty and love."  Both Deano, and Andreas, up wherever you are now, thank you, and Godspeed.  You gave us plenty of flashes of beauty and love to celebrate, to remember you by, and to carry on your work and your legacy...

V