Insights

Does SAP HANA threaten the first serious (hardware) paradigm shift since the IBM PC?

10 Nov 2011 Business Intelligence (BI), ERP, HANA, In-Memory, Consumer Business

If you're in the SAP world, you've heard of SAP HANA, and you know it's fast; but have you considered that it will almost certainly change the world as much, or more, as IBM's original PC did?

Thinking about the impact that HANA will have in the next few years, I was going to title this blog 'Back to the Future', in the hope that Robert Zemeckis and Steven Spielberg wouldn't mind me borrowing their most brilliant film title in return for some free advertising. My original subtitle has now become the main one after I realised that 'Back to the Future' has been so used and abused, in all sorts of weaker and clichéd contexts, that its communication powers have now been lost. Also, John Appleby said, 'it's tacky'.

I'm now going to have to explore what will happen with the commercial availability of In-Memory appliances small enough to go under a desk, without all the fun stuff referring to my crazy haired namesake travelling at insane speeds in inappropriate circumstances and driving a tacky aluminium car with a flux capacitor in it.  I'll also have to put in the effort to think up proper titles for the next two blogs in this series on HANA, and not just 'parts 2 & 3'.

So here goes...

Why SAP HANA is the new 'personal' computer

Ideally, a management consultant should never admit their age, but I'm going to fess up here, so you know where I sit on the information technology timeline - I'm over forty years old and have been exposed to nearly every generation of computing technology since mechanical calculators were all the rage.  For me, SAP HANA isn't really new at all; it's simply the next evolutionary stage of what has been a series of long waves in information technology, and it reminds me strongly of past changes as well as making me think of what will happen in the distant future.

When I first saw how small a HANA box can be, I was instantly reminded of being a teenager back in the eighties; and I'm not talking here about being ignored by girls as the cool kids danced in smoky school discos to Japan and Duran Duran.

I was transported back to a time when 'personal' computing meant the challenge of squeezing a version of Dungeons and Dragons into the Commodore Vic 20's 3.5k of usable RAM; of trying to get the dynamics of a Moon-Lander game working properly on the same machine; and of building a Millenium Falcon laser battle game over a weekend on the Sharp MZ80K (the letter 'H' makes a great Tie-fighter if you don't have bit-mapped graphics available to you).

This was a time when proper computers were room-sized or even bigger, typically needed water cooling and you could see racks of tape drives and enormous Winchester hard drives spinning in their hermetically sealed cocoons; a time when serious computing tasks, such as those my father worked on, had to be set up with large overnight batch-loads (sound familiar?) and were typically run on time-shared machines in the US, because Heriot-Watt University didn't have big enough or fast enough hardware.

There were other micro-computers kicking around, such as the Apple II, which was powerful enough to do tasks such as sorting out our school's timetable.  When visiting Heriot-Watt one day, I saw that someone had brought one of the first Apple Macs from the States, and had sat it down in a research lab like an artefact from the future.  I played around with its dinky little mouse pointer and thought it was utterly fascinating - I could see how people thought this might be a world changing technology and fully expected it to become so.

But in the end, not enough people bought one; certainly no one did in my circle of friends - with their various Amigas, Commodores and BBC machines, all utterly incompatible with each other.  There was, however, a nascent technology that was quietly but relentlessly changing the world that I didn't see coming at all in those early years, although I guess others did. 

By 1984, when the Mac came out, I had mostly given up interest in personal computers and was focussed on buying HiFi equipment to play Led Zeppelin on; honing my climbing skills on mountains like this and discovering that it's not always the 'cool kids' that get the girls, so I was far too busy to witness exactly how the IBM PC snuck up and stole the 'Personal Computer' market from all the other contenders.  There were, however, plenty of clues around that with hindsight are glaringly obvious...

Why is SAP HANA the new PC, and not the new Apple Mac, BBC B or Commodore 64?

This, for me, is where it gets really interesting.  The Commodore 64 was a great little machine; the Acorn BBC B was ahead of its time (Acorn RISC based computing continues to live on in the ARM architecture underpinning nearly all the mobile phones in the world) and Apple was doing amazing things with personal computers, as I described above.  They had clearly managed to sell hardware into schools, and had a very cool product in the Mac; but it didn't change the world quite as Apple had expected it to when it launched with a famous, Ridley Scott directed, advertising campaign suggesting the Mac would smash down 'Big IT' (Read, IBM).

To understand how SAP HANA will turn the information technology world upside down, what has to be asked now is why did all the other 'personal' computers fail and the IBM PC architecture go on to dominate computing in the way it has.  Why do I believe that SAP HANA threatens to do the same as the PC did back in the eighties and early nineties, and create a paradigm shift that will rock all previous assumptions about the direction of the information technology world?

In my opinion, there are many key factors that represent important similarities between the development and marketing of both the SAP HANA technology and the earlier IBM PC one; and these will lead to HANA ultimately setting standards for in-memory computing that will eventually be adopted across the industry.  In no particular order, these key factors are:

• Open, but well defined standards, supported by multiple vendors
• 'Off-the-shelf' hardware
• Flexible and scalable architecture
• Trusted corporate brands behind the hardware and software
• Large existing customer base for the same vendors of hardware and software
• Personal scale in an industrial strength machine
• A focus from vendors on delivering business value first, and fancy packaging second
• A focus from vendors on engineering excellence in the product
• A willingness to allow 3rd parties to develop killer-applications on the platform
  " etc.

I could go on, but I believe SAP HANA will take off because it's going fundamentally to hit the same sweet-spot of huge corporate demand that drove the PC's takeover of most of the world.

Hitting the sweet-spot

For me, there are a few distinguishing features of such a 'sweet-spot' for HANA:

• Relatively simple functional requirements currently held back by current IT processes and technology as Mark Chalfen's blog (Can SAP play a part in stimulating the global economy?) describes
• Business requirements that can be ring-fenced and taken away from the ownership of 'Big-IT'.
• Key enabling technologies, such as Business Objects Data Services and Sybase Replication Server, which allow localised In-Memory solutions to be integrated into the corporate IT infrastructure, as Herman Ohlhoff's blog (DS 4.0 - A quiet revolution that will change Data Warehousing as we know it) outlines.
• Frustrated, but heavy-hitting corporate users,that have some interest in new technology.
• A willingness on the part of those corporate users to begin to judge business value ahead of risk-avoidance, as John Appleby's blog (SAP HANA and innovation hype cycle management) here ably outlines.
• Discretionary budgets that are big enough to simply buy the solution on a 'let's just give this a go' basis.

When a solution like SAP HANA meets corporate demand of this nature, you can reasonably expect changes as big, or bigger, than the PC caused from the mid-eighties until very recently; and I believe these changes will be very dramatic. 

For example, IBM's Phil Allam, with whom I'm proud to be delivering a live HANA demonstration (Building Real-Time Analytics Apps with HANA) at the UK & Ireland SAP User Group conference in Birmingham, brought a PC into British Coal in the eighties with a relatively simple financial model on it.  The next thing that happened was 30 FTEs had to find other jobs, because they were the people that had previously spent days and days building the same model in the existing IT solution and then ensuring that data was dragged laboriously through it.

My uncle was the works director of a company called Peak Engineering in the Midlands when the finance director surprised the rest of the management team by bringing a PC into his office to 'do his own thing with financial planning', when the other directors were working hard to implement an early ERP system.  One of the things that Peak did really well was production cost control, and I don't think it's a coincidence that such visionary thinking on the part of their FD was part of why Peak Engineering not only survived the maelstrom in UK manufacturing during the eighties, but thrived when other companies closed down.

As a student on summer placement in 1987, I nearly tore my hair out when an IT department said they couldn't give me some data 'because the extract program would take too long to code'.  Fortunately, colleagues had already pulled the data into a PC with dBase running on it, so we could completely bypass the central IT guys and get the information quickly into a marketing report.

When my father was writing software in the 60s and 70s that ultimately (albeit indirectly) helped to stop the Space Shuttle main engines from blowing up, he needed the largest mainframe class machines in the US to run the code overnight. When he ported it to the 486 PC to make use of its maths co-processor, the model ran locally in less than 4 minutes, so clients such as BP, that are much poorer than NASA, could use it on a large-scale to optimise the running of big refinery feed pumps.

When I first used CAD and Finite Element Analysis, during a summer placement in 1990, the minimum hardware you needed to run the FEA models was something like a DEC MicroVax, but where are DEC and VMS, now that PC based solutions have properly grown up?

So with SAP HANA beginning to hit the In-Memory Sweet-Spot, what will play the role of the Dec MicroVax in this story?

As described above, the IBM PC architecture began in the eighties to encroach upon, and then dominate, sectors that had previously been supported by much heavier hardware and more laborious IT processes; until whole classes of hardware and software became obsolete and rolled over under the relentless pressure of Moore's law and open PC-based standards.  Along the way it has grown up and out until there is now no such thing as a 'mini-computer' and mainframe class computing (read, the cloud) is being done on remote server farms built with tens of thousands of PCs.

And I believe that HANA will ultimately do exactly the same and we need to imagine what a world built 'In-Memory' might eventually look like. In the next two blogs I am writing, I'm going to explore how HANA will do this; some of the specific business scenarios that will allow Bluefin Solutions to help it to do so in the short and medium term and some thoughts about what the resulting long-term paradigm shift might look like.

Because I'm explicitly not going to go back to the future, what I won't do in these future blogs is describe how HANA will dump a steaming pile of manure all over the IT equivalent of Biff Tannen; I'll simply try the more prosaic task of identifying who, or what, will play the role of the DEC MicroVax in the story of why HANA is the new IBM PC.  I hope you enjoy them.