Category Archives: HANA

The evolution of Analytics with SAP HANA

I was recently sent a collection of analytical business questions by a customer. I can’t share the details of them of course, but I was surprised by the level of sophistication in the questions – they included product, time, demographic and geospatial dimensions.

The analytics team was trying to get answers to these questions for their executive team and they were encountering time-outs trying to get the data out of their existing data warehouse. The vendor will go unnamed but it is a traditional disk-based EDW platform.

The analytics team had heard we had built a SAP HANA system, which included some of the elements around which they were asking questions, and were interested to see if we could help answer them using SAP HANA.

In the meantime, I was reading Holger Mueller‘s interesting blog on Why is analytics so hard? Or: The holy grail. It’s an interesting read, come back here once you’ve read it. And following this, I had a bunch of thoughts:

What did we build in HANA?

Again I can’t share the specifics, but I can tell you what we did, generally. We loaded Point of Sale data (not a retail customer) and a collection of master data like Customer, Supplier, and also a bunch of reference data like Geospatial, Demographic.

What’s really cool about doing this in HANA is that we can use as many CPUs as we have available to process a question, and we have 160 cores in my test box. This means that we can answer a question like “Tell me about the average sale price by customer, and slice it by the age of the customer at time of transaction, and group it by political leaning”. And HANA returns the answer in 2-3 seconds. Any question you like. Keep asking questions.

Now we used both SAP’s Lumira tool and Tableau on top of this to do visualizations. Lumira is a little quicker, Tableau has a lot more features. And I came to some conclusions:

1) HANA brings structured analytics into Holger’s innovation “Phase 3 – A business user can – with appropriate, but affordable training – use the innovation”

You have to be able to formulate questions, and use a tool like Tableau/Lumira. I learnt how to use Tableau in a few hours with no training, and Lumira in a few minutes. But, you have to know what questions you want to ask, and they have to have meaning.

More specifically, we could answer some of the questions using the model we built, and he wouldn’t have timeout problems.

2) Data quality problems always lurk under the surface

It won’t surprise anyone who has worked in data warehousing, but data is a big challenge. It’s not possible to easily answer some of his questions because their category hierarchy doesn’t allow for it. They have categories A-E and his question wants to know about category F. Category F has business meaning but their system hasn’t been updated to know that category F exists.

This requires an update to how they process master data and assign categories to transactions. We can actually do this really easily in HANA. For instance, we could use publicly available reference data like SIC codes to process this and then reprocess the transactions. Because we never need to aggregate with HANA, we only have to do this once and we’re done.

3) The structured data we built so far is not enough

There are data elements that we didn’t include in the initial model which means that some of the questions being asked can’t be answered yet. But also, some of the business questions are sophisticated and based on the latest trends, so the model hasn’t evolved.

We can add this stuff into the structured model easily enough, or with Tableau you can join a HANA model with a Tableau model so you can load that stuff into your own Tableau software and then do analysis. But suddenly we’re in Innovation “Phase 2 – Through tools more trained professionals in the relevant technology can make the innovation happen”.

And with HANA, we have to be careful with our data model if we want sub-second performance on billions of records, which may push us into Innovation “Phase 1 – Only experts can apply the technique to make the innovation happen” if the structure change is substantial.

Conclusion – the Layered Structured Architecture

And this is where my mind is headed: we have to classify how we want information to be available and what sorts of extensions to the analytics model different types of people can do.

For example, Phase 3 users will quickly find they can do quite advanced analytics. For example with Tableau you can easily join against Outcode to do geospatial analysis. And it performs great when joined against customer data in HANA.

But if you want to include new transaction categories (data) then you’re in Phase 1, which means you need a process to regularly update your structured model to include the new things people are thinking about.

Good business people will keep asking for more, harder questions. I think with HANA we have a platform which facilitates this, rather than handcuffing us.

SAP HANA: disrupting 3 markets, but does SAP have the stones to disrupt itself?

I’ve been immersed in SAP HANA ever since I first heard about it, when Hasso Plattner talked about it in May of 2010. SAP HANA is an in-memory database that computes thousands of times faster than traditional database products.

There are two business lessons that I learnt in this time: first, amazing products are products that disrupt markets and create new markets. Think of the Dyson vacuum cleaner or the iPhone, for example. Second, as Steve Jobs said, “If you don’t cannibalize yourself, someone else will.”

What’s fascinating about SAP HANA is that it has the capability to disrupt four market categories. Even SAP themselves don’t see the full potential of HANA – and that’s fine, because the most disruptive technologies don’t always know their purpose in advance. Let’s get into the detail of these categories and what it means.

1) Database Market

This is the most obvious – SAP HANA is a database much like Oracle, but it runs entirely in-memory. This means it is 1000x faster for many operations and scales linearly. As SAP CTO Vishal Sikka told me this week – “If it runs in 100 seconds with 1 core, it can run in 1 second with 100 cores. That is the beauty of HANA”.

This matters because there are things you can’t do with Oracle that HANA can do. Cancer genome sequencing in seconds. Real-time monitoring of oil drills. Providing offers to customers at a till based on their basket and historic purchases.

In addition, SAP HANA can now be used as the database for all of SAP’s existing Business Suite install-base – 89,000 customers, and it’s a lift-and-shift simple process to move off Oracle. Once you’re on HANA, you can start to take advantage of things you couldn’t do before.

As of now, SAP is the fastest growing database vendor in the world, and overtook Teradata as #4 in June 2012. They booked $800m of database revenue in 2012 (see slide 37). SAP HANA is disrupting the database market.

2) Enterprise Hardware Market

Most Enterprise Software customers run software on expensive mainframe-style hardware. Individual boxes can run into the millions. So SAP built their HANA database on high-end commodity Intel x86 hardware. HANA will run on your laptop (if you have enough memory) and it’s supported by a number of the usual hardware vendors.

SAP expected to disrupt the hardware market, as well as the expensive enterprise storage market, which is based around selling outdated spinning chunks of metal. The market already moved onto solid-state storage and it’s known to be reliable (how many iPhones, iPads and Samsung Galaxies do you see?).

Unfortunately, the hardware vendors like IBM and HP saw SAP HANA as a means to sell yet more spinning disks, and so created hardware configurations with even more of them. SAP got annoyed and invested $50m in Violin Memory, a start-up that has a solution to this. No vendor has yet produced a SAP HANA configuration based on Violin, though I hear SAP bought a ton of them. [update, a colleague tells me that Fujitsu went live with a customer on a Violin appliance, but it's not officially certified hardware]

So this year I expect further disruption to the hardware market, perhaps with a new entry into the x64 Server market that knows how to do premium commodity. Lenovo, perhaps?

3) Services Market

One thing is for sure: SAP wouldn’t be where they are today without Accenture and IBM Services. Those guys pimped SAP R/3 in the 90s and made SAP the success that it is. In the process of this, they created an even bigger market for themselves – 6-8x larger than the Enterprise Software market.

This really annoys SAP co-founder Hasso Plattner and to quote SAP CTO Vishal Sikka: “we have this strong strategic need to not have the partners come in to implement HANA. If that happens, then we have failed, and Hasso told me that.”

I talked to Hasso about this last week and he was unrepentant: SAP have to solve the services market problem. I told him he had got it wrong and SAP as an SI was just the same as Accenture and IBM and the ratio is not decreasing for HANA projects. If SAP wants to disrupt the consulting market then it needs to do so by facilitation – like the Violin model.

Services disruption is of particular interest for me and I’m always pushing my consulting teams to innovate. I noticed that on a recent deal, the ratio was inverted: 10% Services, 30% Hardware and 60% Software. Of course, we will do more services work once the software is installed, but it gives you a flavor of how the services market might be disrupted, if we can scale our models.

Interestingly, SAP Global Head of Sales Rob Enslin just got put in charge of SAP Services, and he’s an ex SAP Basis consultant. I talked to him last week and he is changing the organization in ways that I believe will support this disruption.

Can SAP disrupt itself?

If you know anything about the Enterprise Software market, you are probably wondering why I missed the elephant in the room: The Cloud. I didn’t, I just left it until last. Here is SAP’s biggest challenge for 2013.

SAP HANA is sold in one of 3 ways. One, you can buy a runtime license for your Business Suite for 15% of your Software Application Value (total of what you paid SAP) per annum. Two, you can buy Enterprise licenses which are bought by the 64GB chunk of memory (minimum 128GB). Third, you can buy 60GB of HANA One on the AWS marketplace for $1 an hour (plus $2.50 an hour for hosting).

Now, Amazon are a very impressive company and they are creating bigger machines – 244GB systems were released last week. As of now, SAP haven’t announced if they will be supporting them and if so, what the cost will be.

However the problem is simple: HANA One doesn’t cannibalize on-premise sales, because you can’t buy 60GB of HANA from SAP on-premise. 120 and 240GB Amazon instances do, however, cannibalize the low-end of SAP HANA on-premise sales and this will have a negative impact on revenue perpetual license revenue.

But the fact is: the world is going cloud. As William Gibson said: “The future is already here — it’s just not very evenly distributed.” – and whilst there are many companies who will run on-premise software for a very long time, companies as big as Pepsi are implementing cloud solutions (they just bought 300k Employee Central licenses from Success Factors).

My view is that in any case, providing a cloud subscription model for HANA can only be a good thing. It dramatically simplifies adoption and – at least for the medium term – there won’t be an option to do cloud HANA for large installations so customers will have to move on-premise at some later stage, and will probably then buy even more HANA. Moreover, Wall Street and thereby the shareholders put a higher value on recurring revenue (Workday is valued at P/S of 41x, whilst SAP is valued at 6x), which should help.

So as for me, there are two crucial questions here:

1) Does the SAP board have the stones to accept a short-term hit in revenue and move to the recurring revenue subscription model for HANA?

2) When will SAP enter the in-memory cloud market? For every $1 they take for HANA One, they are giving away $2.50 to Amazon for hosting. This is money left on the table.

2013 is shaping up to be a very interesting year.

The SAP HANA Career Guide – Part 5, SAP HANA BW Consultant

Hopefully you have enjoyed the SAP HANA Career Guide so far. This piece focusses in on the SAP HANA BW Consultant. These guys are responsible for upgrading, migrating and HANA-enabling existing information on the SAP BW Enterprise Data Warehouse, as well as the creation of new Data Warehouse solutions.

Where do SAP HANA BW Consultants come from?

Well this one is easy! Any existing SAP BW Consultants – especially those who are more business focussed and don’t have the heavy hitting SQL skills that would make them great SAP HANA Performance Consultants – can make excellent SAP HANA BW Consultants.
This is largely because running SAP BW on HANA is broadly similar to running SAP BW on any other database. The modelling principles, business object principles and key considerations for things like stock or currency conversions remain exactly the same. So if you’re an existing SAP BW consultant then look no further.

What does HANA Thinking mean with BW on HANA?

There are a few important changes worth thinking about. The first are architectural. BW on HANA requires fewer objects – you lose Indexes, Aggregates and some of InfoCubes as well as being able to lose certain types of DSO and Master Data objects.

This means simplification of both the number of objects and with that, data loads and query management. And that brings with it a simplification of project design, methodology, reduction in load times and testing times. It completely changes the way that BW projects are run – reducing project timelines and increasing time-to-value.

How do I cross-train to SAP BW on HANA?

The SAP Customer Solutions Adoption team have produced an excellent course for experienced BW people called “LSA++ – THE LAYERED SCALABLE ARCHITECTURE FOR BW ON HANA“. As is the CSA style, this is designed for those who already have great BW skills and need the HANA specific stuff.

Here, they explain the difference in thinking between HANA and any other RDBMS, and what that means to architecture, design and the practicalities of Enterprise Data Warehouses.

What Classroom Education is available for SAP BW on HANA?

The classroom training is really limited in this example. SAP Education have a course called TZBWHA: SAP BW on SAP HANA – but it is twice the price and contains half the content of the SAP course. I understand that a new course is being written as we speak – hopefully it will contain the right content. In the meantime, I don’t recommend this course.

Where can I go to ask questions?

As before, here are two great places for this. First there are the SCN SAP HANA and in-Memory forums, where you can ask technical questions about all things SAP HANA. Response times are excellent.

Second, you can go to the Experience SAP HANA Discussion area, where there is a similar focus on assistance.

The SAP HANA Career Guide – Part 4, SAP HANA Operations Consultant

Hopefully you have enjoyed the SAP HANA Career Guide so far. This piece focusses in on the SAP HANA Operations Consultant – which would have been called SAP Basis for regular SAP systems. I’ve always hated this term, and thought it was time for a new one, and Operations is all about getting things running and keeping them running – efficiently.

Where do SAP HANA Operations Consultants come from?

I think the reality is that may of them will come from SAP Basis but there are some important things to note. First, is that SAP HANA only runs on SuSe Linux, so knowledge of other platforms (Windows, UNIX) is only tangentially relevant.

What is relevant is design of technical architecture – although SAP HANA solutions are created from building-block principles and so there are a limited number of possible configurations. Knowledge of High Availability and Disaster Recovery principles are a must, as most SAP HANA implementations require this.

To add to this, a working knowledge of Linux administration, script writing (bash, awk, Python), X-Windows, ELILO as well as networks: all High Availability SAP HANA appliances must have 10 Gigabit Ethernet, for example. Plus, if you have IBM hardware, a knowledge of the GPFS clustered filesystem is a must.

What does day-to-day administration of SAP HANA look like?

Once SAP HANA is set up, it requires remarkably little attention. New nodes are installed with a single command. If you add or remove hardware, one thing you do need to do is redistribute tables between nodes, but this is also quite straightforward.

No optimisation, re-indexing, indexes, aggregates or other elements are required in regular operations so the DBA overhead is much lower than other databases.

How do I find out more about SAP HANA Operations?

To be honest, the SAP HANA Master Guide provides all you need to know and there is a Technical Operations Manual available. If you are already a DBA or SAP Basis consultant with the skills listed above, I recommend you dive right in.

One challenge is getting the SAP HANA software for testing purposes and I hope to have some good news on that this year! If you are a SAP Services Partner then you are able to get the software at a good price as a Test & Demo license.

What Classroom Education is available for SAP HANA Operations?

There is a specific SAP HANA Operations course called TZH200, which may be worth taking if you enjoy learning in a classroom environment. This leads to a certification qualification possible called SAP HANA Certified Technology Associate.

What about running SAP on HANA?

If you run SAP on HANA then you will also need to know SAP Basis – the fundamentals of which are well documented.

In this case, you are probably interested in migrating SAP systems from some other database like Oracle onto HANA, and in this case you do need some special experience. SAP mandate (and I also recommend wholeheartedly) becoming a SAP Migration Certified Consultant, which is a significant investment. If you do not have this certificate then the systems you migrate will not be fully supported by SAP.

If you are migration certified and you have learnt the above material and familiarised yourself with SAP HANA, table partitioning, row- and columnar-stores, the way that HANA manages deltas etc. then you are ready to do SAP HANA Migrations and could call yourself a SAP HANA Migration Consultant. I haven’t created a separate page for that because I believe it is the same core type of person.

Where can I go to ask questions?

As before, here are two great places for this. First there are the SCN SAP HANA and in-Memory forums, where you can ask technical questions about all things SAP HANA. Response times are excellent.

Second, you can go to the Experience SAP HANA Discussion area, where there is a similar focus on assistance.

The SAP HANA Career Guide – Part 3, SAP HANA Performance Consultant

Hopefully you have enjoyed the SAP HANA Career Guide so far. This piece focusses in on the SAP HANA Performance Consultant.

In the early implementations of SAP HANA, this was by far the most popular type of resource. SAP HANA Enterprise comes with a set of tools including SAP HANA Studio, which contains the SAP HANA Modeller and the SAPScript programming language (which is similar to PL-SQL).

The SAP HANA Performance Consultant takes requirements and builds data models, including the virtual Analytical views and Calculation Views that make SAP HANA special, and builds the SQLScript and CE Function programming code to meet the needs.

Where do SAP HANA Performance Consultants come from?

Whilst building simple SAP HANA models is something that almost anyone with knowledge of Microsoft Access can do, the SQLScript language and CE Functions are technical languages that require a sound programming understanding. Those familiar with programming stored procedures in RDBMS systems like DB2, Oracle PL/SQL and Microsoft Procedural SQL will find themselves at home quickly.

Similarly those familiar with the SAP BW Data Warehouse may find themselves out of there depth here. Those SAP BW consultants familiar with writing complex transformation and update rule code in the ABAP and OpenSQL programming languages may find SAP HANA Enterprise comes naturally – especially those with a technical background and degree. Those who are more business focussed and less technical would be best advised to focus on the SAP BW on HANA consultant.

How do you cross-train from PL-SQL to SQLScript?

The programming languages are fairly similar and any SQL developer will be able to familiarise themselves very quickly by referring to the SQLScript Guide. Note that this guide is updated every 6 months with major amendments, when new releases of SAP HANA are made.

How do I get hands-on with SAP HANA?

The best way to cross-train is to get hands on and build data models. Thankfully the lovely folks at SAP have made this really easy. There is a 30-day free developer version of HANA in the cloud available in the HANA developer center. After that, you pay by usage of the Amazon AWS HANA system – the SAP HANA software itself is free to use for test purposes.

In addition, the SAP HANA Distinguished Engineers are building out a collection of fantastic learning videos that take you through each of the SAP HANA concepts and get you up and running fast. This will be called the HANA Academy and is coming soon – I will post details as soon as they are available. In the meantime there is a YouTube video with the content.

What classroom training is available?

SAP offers a good basic training guide called HA300 as a 5-day course. If you learn best in a classroom environment and can afford the €2500 cost (plus expenses) and time out, then this might be a good option.

Be aware that this course is typically out of date: SAP HANA moves very quickly and classroom education struggles to keep up.

Where can I go to ask questions?

There are two great places for this. First there are the SCN SAP HANA and in-Memory forums, where you can ask technical questions about all things SAP HANA. Response times are excellent.

Second, you can go to the Experience SAP HANA Discussion area, where there is a similar focus on assistance.

The SAP HANA Career Guide – Part 2, SAP HANA Business Consultant

Hopefully you have enjoyed the SAP HANA Career Guide so far, which kicked off with Part 1, Overview. This piece focusses in on the first specialty: the business consultant.

The Business Consultant

Business consultants analyse the needs of the business and create a strategy to transform businesses, or line of businesses. The essence and principle of this doesn’t change with SAP HANA – indeed it is much of the same. They listen to the needs of the business and the ways in which it must change, and then apply technology concepts against that to create a technology strategy.

For instance I have a customer who has a problem with fraud prevention. The business consultants came up with a solution that enables the business to reduce customer fraud.

Why does SAP HANA change this?

That’s just it: SAP HANA doesn’t change the needs of the business consultant. Instead, it changes the envelope. Let me illustrate how the technology changes the envelope.

I have a customer where we used SAP HANA to accelerate sales order reports within ERP. The approach was to read a bunch of sales order headers, get detail from within, cross reference them against various attributes and exclude a bunch of orders, to produce a report. This requires between 10-20 thousand questions, that SAP asks and takes 30-60 minutes to produce a report.

Now, SAP HANA performs the same as any other database to answer 10-20k small questions. But, with some small changes, we changed 10-20k small questions into one huge question. What happens now? SAP HANA responds in a few short seconds.

Why does that matter?

The answer is, in itself, it doesn’t. However, now we can give these repots to sales execs on the road and they can access them on a mobile device in seconds, giving information about past spend, profitability and other key elements. But even that doesn’t pull the real power of SAP HANA.

The real power is when we move this thinking to whats happening in the moment – let’s take automotive as an example. We can collaborate with a customer to create a quotation with them – price that quotation against very complex pricing structures that exist in complex business models. Even calculating margin on the fly against a car with 10,000 parts. Looking at upstream supply chain visibility to see that removing an option for an automatic gearbox changes the delivery date from 4 months to 1 month. Discounting on the fly based on available stock and the desire to sell particular options.

In this case, the customer experience becomes collaborative and communicative and you can close the deal in the moment, rather than having to come back with a quote the next day and an estimate for delivery 3 weeks later. It’s real, and customers will buy it.

SAP HANA Use Cases

The first thing the business consultant needs to do is to read about SAP HANA use cases and consume them for their industry. They are available for public consumption at Experience SAP HANA and this will begin to cultivate a HANA State of Mind.

The HANA State of Mind

I have written about this before and I see this change in consultants who get immersed in SAP HANA. Once you see the capabilities, you will be able to apply “HANA Thinking” to everyday live. You will see the business possibilities where producing a particular report can reduce costs by millions of dollars a month. Only you thought it wasn’t possible.

Education and Training

Here’s the kicker with SAP HANA Business Consultants – I’m not sure that it can be taught. You have to combine existing Line of Business and Industry expertise with the knowledge of how SAP HANA can disrupt businesses. If you’re good at business consulting already, all you have to do is to understand how SAP HANA can help.

Perhaps I’m wrong here and I’d like to be challenged. Perhaps there is a “Power of SAP HANA” set of webinars, videos or instructional content. What do you think? Is it just a question of repeating a few business scenarios where SAP HANA makes a difference, and business consultants will just “get it”? Let me know.

The SAP HANA Career Guide – Part 1, Overview

SAP HANA is one of the fastest growing software technologies ever. It was released mid-2011 and sold $250m in the first year. In 2012, financial analysts expect $500m+ and it is expected to be a $1bn+ market in 2013. In software services terms, this is at least a $4bn market, next year.

It’s hardly surprising, therefore, that there is a huge interest in training, education and certification right now and there isn’t a good guide to what makes a good SAP HANA consultant. So, in a 8 part series special, I am going to lay out what types of consultant exist, where to get educational resources and how to get ahead in getting a job in the SAP HANA market.

SAP HANA creates a new category of consulting

It’s really important to note that there isn’t just one type of “SAP HANA Consultant”, whatever someone tries to tell you. SAP HANA, like any other technology, has a number of sub areas. My advice: first, understand the different categories and then decide – based on your experience and interest – which one of these you will be best at specialising in, and learn that subject matter.

This guide will help you understand which category you fit in, and where to find the resources to become an expert.

SAP HANA Distinguished Engineer Program

Whatever type of consultant you are looking to be, I recommend looking at the SAP HANA Distinguished Engineer Program. I’m on the council, so I would say that – but the program will support you in your learning, and then recognise you for your achievements, knowledge and community efforts. Read the FAQ for more details.

Let’s get on and discuss the different categories of SAP HANA consultant directly.

1. SAP HANA in-memory Business Consultant

Business consultants understand one or more industry verticals, and typically a number of lines of business, with a specialty – for example Retail Sales, or Utilities Supply Chain. The SAP HANA in-memory Business Consultant also understands how in-memory technology technologies can disrupt businesses and gets how to apply the technology concepts to business scenarios.

It’s fair to say that this consumes a good number of what used to be described “functional consultants”. In some instances, like the Finance Line of Business, their skills may be applied cross-industry.

Click Here to link to the main article

2. SAP HANA Performance Consultant

SAP HANA leader Steve Lucas and I coined this term earlier in the year because it is very emotive. The SAP HANA Performance Consultant takes the concepts developed by the Business Consultant and creates solutions, architectures and designs using SAP HANA Enterprise. Typically these include using the SAP HANA Modeller tool to create models.

Click Here to link to the main article

3. SAP HANA Operations Consultant

This is what in traditional terms used to be called SAP Basis, but I always hated that term and was glad that a new term could be coined! SAP HANA Operations Consultants understand technical architecture, Linux, how to install SAP HANA or migrate systems, and also have some understanding of SAP Basis.

Click Here to link to the main article

4. SAP HANA BW Consultant

The SAP HANA BW Consultant is fairly similar to a regular BW technical consultant but there are some specific skills about architecting, re-architecting and modelling BW solutions within SAP HANA that are essential to know.

5. SAP HANA Application Developer

SAP HANA requires a change in the way in which you develop applications, so whatever type of apps you are looking to build – be it ABAP applications using the HANA database, or mobile applications using the HANA XS Application Services layer, you will need to understand your existing development platform, and in-memory computing concepts. The SAP HANA Application Developer combines these skills to produce high-performance apps.

6. SAP HANA Security Consultant

This is a niche area but one we can’t do without. Security design, access control and security models are different because of the way that SAP HANA is designed and you need an appreciation of business security concerns, access, audit and compliance as well as SAP HANA to do this well.

7. SAP HANA Project Manager

I didn’t add this category into the original blog post but I’ve decided to add it: managing SAP HANA projects requires the usual PRINCE2-style project management experience, but combined with some elements of agile methodologies, but more importantly taking into account how SAP HANA changes project life-cycles.

Final Words

Hopefully you understand from this article which of the categories fits you best. I think it is safe to say that, provided you have the background, skills, and desire to be an awesome HANA Consultant, any of these could be a lucrative career choice.

In the following 7 articles, I will discuss each of these roles in detail, along with any relevant certification programs, training material and other self-help locations, so you can pursue your SAP HANA career of choice. Good Luck!

Credits

As usual this post wasn’t possible without the help of others. In particular, thanks have to go to the rest of my HANA Distinguished Engineer Council members: Michael Eacrett, David Hull, Harald Reiter, Jon Reed and Vijay Vijaysankar. But also to nearly everyone else I spent time with over the last 5 years.

Inside Oracle Exalytics: Oracle shows its fear of SAP’s in-memory strategy

Oracle released its new in-memory analytics engine, Exalytics, at Oracle Open World 2011 this week. Media has been in a frenzy for the last 4 days and there are already over 1/4 million Google hits for Exalytics. It’s actually interesting though, as to why Oracle would release such a product.

First, Oracle’s Ellison publicly dismissed in-memory computing just 6 months ago. He claimed SAP were on drugs, building out an in-memory database, as referenced in this PC World article:

“Get me the name of their pharmacist,” Ellison said at the time. “I mean, I know a lot about in-memory databases. In fact, we have the leading in-memory database, TimesTen. This is nonsense. There is no in-memory technology anywhere near ready to take the place of a relational database. It’s a complete fantasy on their part.”

Second, Oracle has its Exadata database appliance – which is designed to serve the needs of both online transactional processing, and analytics. With Exadata, you are supposed to have all the appliance you need. Here’s what Ellison said when he unveiled Exadata:

“It is the first database machine that does online transaction processing. All the other machines: Teradata, Netezza etc. are designed just for data warehousing. Oracle Exadata version 1 was designed just for data warehousing. This is the first time a database machine has been able to do both data warehousing and online transaction processing.”

What does an Exalytics appliance look like?

The reason why Oracle were able to produce the Exalytics appliance so quickly is because it’s mostly off-the shelf hardware and software for them:

  • Sun Fire X4470M2 server with 40 cores,  1TB RAM, 6x 600GB SAS drives
  • TimesTen columnar database store (columnar feature is new)
  • Essbase OLAP database (optimised for in-memory)
  • Oracle BI Suite

So Exalytics is nothing new?Not really. It is a bunch of existing Oracle tech, optimised for in-memory, although they don’t say how, apart from the columnar TimesTen database, which will have 5-10x compression compared to the row TimesTen store.However it is neatly packaged up into an appliance which is easily purchased, and one which will probably provide good performance benefits compared to even a regular Exadata appliance.

Why is Oracle afraid of SAP?

Oracle have no need to build this appliance. It’s architecturally complex and they would make more money out of selling the premium Exadata appliance. So why did they? For my money, because Larry wanted to show SAP that he could also build an in-memory analytics appliance in a short period of time.

And this shows for the first time – in the wider market context – that Oracle might be afraid that SAP’s strategy of in-memory computing with its HANA appliance, is the right strategy. Unfortunately Oracle have missed the point.

What is the point and why does Exalytics miss it?

Oracle Exalytics and SAP HANA solve the same high-level business problem today – poorly performing reports. I’d bet if you loaded the same data into them, they are probably quite similar and would provide the same value to the customer. The problem for Oracle is: that’s not the point of SAP HANA.

SAP HANA is a simple and elegant solution that – whilst not quite there today – will, in its roadmap of future versions provide the following:

  • Excellent performance for transactional and reporting systems on the fly
  • Integrated planning engine with no separate store like Essbase
  • Elegantly architected with row and column stores and the benefits of both
  • Generic RDBMS capabilities for all business applications and platforms
Exalytics on the other hand is a bolt-on mixed-technology appliance. For Oracle to really compete with SAP, they need to throw out Exadata and Exalytics and build an in-memory RDBMS appliance that can do what both of those appliances do in one. They have the brains to do so – but will they? We will see.
Update 14/10/2011:
Thanks to David Hull for a couple of clarifications!
1) I don’t make it clear that the Exalytics appliance which is a Sun Fire X4470M2 is an Intel-based server like SAP HANA, rather than using Oracle’s SPARC technology that it acquired from Sun. There is a nice irony in this. I’m actually wondering why they use the Intel tech. It’s cheaper – but most of the components seem to run on SPARC as well.
2) TimesTen is not a columnar store – it’s still a row-based store, but is optimised using columnar compression – which gives some of the compression benefits of a columnar store, but without all the performance benefits. I missed the distinction here.

Oracle & SAP: sticks and stones may break my bones

… but as the saying goes, words will never hurt me. I bet if a SAP HANA or Oracle Exalytics box landed on your foot, it would hurt though.

SAP created a new class of computing this year – it brought together high-end commodity hardware, new software and large amounts of RAM. The era of in-memory computing is on its way. I have, at times, been critical of SAP’s overmarketing of this at a time when the product was not ready for market. Marketing is all about making products easy to sell, and hype extends the sales cycle. That said, I also believe that in time, SAP HANA will be seen as an inflection point in computing. It may or may not be the dominant supplier, but it got there with the right appliance at the right time.

So it’s no surprise that Oracle entered the market soon afterwards, with its Exalytics in-memory appliance. On the face of it, SAP HANA and Oracle Exalytics are very similar and they definitely reach out to the same market – those people who want mid-volume (1-5TB of traditional data mart), extremely high performing, complex analytics.

What is also clear when you get into the detail, is that SAP have architected something new and very interesting – HANA has the potential to be an application platform, database and appliance that underpins all of SAP’s enterprise software, and also potentially software from other vendors.

By contrast Oracle Exadata is (today) a patchwork quilt of existing software, shoehorned into an in-memory appliance. Oracle has its work cut out for it building out a more cohesive and simple in-memory appliance. But note that Oracle does have an existing stable and mature RDBMS. SAP does not, and it must go through that product maturity cycle with HANA.

So why the willy waving?

With this in mind, why on earth have SAP reacted the way they did, to the Oracle news. It’s traditional for Hasso Plattner and Larry Ellison to abuse each other a bit (Hasso famously mooned Larry), and it’s definitely business-as-usual for Oracle to name call in public, like Hurd and Ellison recently. But remember Plattner is not a SAP employee and can do whatever he wants.

Normally however SAP are more demure. Not so this week, and it all kicked off with Bob Evans (a paid journalist by SAP) who wrote on Forbes: Oracle President Safra Catz Suffers SAP Hallucinations. This stuff reads like wartime propaganda. Check it for yourself. That was Monday and Bob is a journalist by trade and prone to sensationalism – check some of his other posts. I can forgive him for his content because he produces what he’s told to do.

But then came Tuesday, and Chief Marketing Officer Jonathan Becher and Head of Technology and Innovation Platform Marketing for SAP Aiaz Kazi took centre stage. They put the simple message out there with his Forbes post: Why SAP HANA is a Better Choice than Oracle Exalytics. This piece has the frosted gilt of Bob Evans’ hand, and I’m guessing he edited it.

Customer-centric: In attempting a retro-IBM style move with their hardware infatuation, Oracle has forgotten the customer. I watched hours of videos and read litanies of blogs but I can’t find answers to simple questions.

Really SAP shouldn’t be throwing stones in glass houses – SAP HANA information can be hard to find, and it is by no means a market-perfect solution. But again I can forgive Jonathan and Aiaz – I think they got a bit carried away with their story, and the editing pen made it sensationalist.

And then Wednesday came. President and Corporate Officer, Global Solutions at SAP, Sanjay Poonen, logged into Forbes and posted his tuppence worth: Oracle’s Exalytics: “Old Wine in New Wineskins”. The propaganda continues with:

CIOs, the choice is yours: sign up for Oracle’s Exa parade and its never-ending set of hardware purchases, licenses and maintenance, and it will increasingly lock-in control over you and your future, with antiquated architectures.

I don’t really know what to say to this.

Why does SAP appear to be threatened by Oracle Exalytics?

So why? Why the parade, the fluffing of feathers and the sharpening of claws? Why the impromptu video of CTO and board member Vishal Sikka? Why?

The only clear reason appears to be investors and share prices. If the market thinks that SAP HANA can be replicated by Oracle in 10 short months then they might think that SAP HANA isn’t so special after-all and the financial community might reduce their expectations of SAP’s stock.

The problem is that I talk to a number of these analysts on a quarterly basis and they are very smart people. They take advice from a number of people in the technology community and crunch numbers for a living. And they live in large-enterprise politics and can see straight through propaganda.

A parting thought

What’s more, they will understand that SAP HANA is (currently) the right way to architect in-memory computing and Oracle Exalytics is not. They will also understand that this is the first battle and the war for dominance in this space has only just started.

And if you’re listening SAP right now – throw away Thursday and Friday’s blogs (who next, McDermott?) and focus on creating clarity of vision of the go-to-market proposition. Because that, is what CIOs buy.

Update

Based on further conversations with Enterprise IT experts, I can confirm another dimension that I hadn’t thought of. The mass of SAP rhetoric has sent parts of industry into a temporary panic. Why, they muse, is SAP so bothered about this? There must be something very damaging in Oracle’s announcement and SAP must be running around to cover it up?

Probably not what the execs at SAP were looking to achieve, unfortunately. By the way those experts seem to have gotten to the same conclusion that we discuss here: Exalogic is not so damaging to SAP, and the reaction is therefore bizarre.

Update 7th October 2011

This post seems to have caused some upset. I thought it was reasonably objective (and it is certainly not my intention to upset people) but this is the response I got from board member Vishal Sikka:

@vsikka: @MatthiasGerisch @applebyj so tons of people ask you to respond, you do, as humbly as possible and without bs, and that’s propaganda?  Wtf.

Vishal’s response is understandable to some degree, given that his YouTube video is indeed candid and quite objective. But again the point I was trying to make was in two dimensions:

1) Why criticise your competitors at all? I work for a SAP Systems Integrator and you don’t see me publicly criticising competitors like Accenture, IBM or Deloitte. Customers will buy my organisation’s services based on the quality of our value proposition alone. Criticising my competitors is not useful and customers are usually distrustful of people that do this.

2) SAP’s board is responsible for its overall outward response.  It doesn’t matter about an individual response – it is the overall picture of articles, interviews and blogs that is painted, including those people that write them. And it is this picture – not an individual video – that appears to be propaganda in this case.

And if SAP’s board isn’t happy with its overall media response, it needs to do something about it.

P.S.: I joked that Thursday’s post on Forbes would be by McDermott. Amusingly, it is, although it is by Tim Clark and entitled SAP’s Bill McDermott: “It Doesn’t Take Two Years to Create a Good Strategy”. Not a dig in Oracle in sight. Phew.

SAP HANA vs Oracle Exalytics – the game is on

So Oracle announced Exalytics yesterday and I’ve spent much of today thinking about it in the back of my mind whilst I did my day job. It’s about the most interesting announcement of the tech world this year and it’s worth discussing why.

What is Oracle Exalytics?

Well it’s a bit like SAP HANA. It’s a big (relatively) commodity piece of hardware with 40 Intel cores and 1TB of main memory. It can store a claimed 5-10TB of compressed analytics data, but beware of false compression claims. On this, you can perform high-speed analytics on medium-size datasets.

Exalytics is a technology mashup of its TimesTen in-memory database technology, BI stack and Essbase OLAP engine. This means that it’s not anything new, but rather existing technology repackaged into an appliance. Not necessarily a bad thing, because it means it should just work.

How does it compare to SAP HANA?

Compared to SAP HANA 1.0 SP02, Oracle Exalytics is almost exactly the same. It is a combination of Hardware and Software packaged into an appliance. The major difference is that Oracle build the whole stack so you can buy an Exalytics appliance directly from Oracle.

However it is only available in one size: 1TB, which is a strength and a weakness. First, SAP can sell anything from a 64GB license (with a minimum 128GB appliance), which at list price is less than €200,000 – all the way up to a monster 4TB appliance in partnership with IBM (and probably other hardware partners, later in the year).

Oracle validate SAP’s in-memory strategy

That’s what’s interesting here: Oracle is validating that SAP has got it right. Larry Ellison usually goes his own way and it is a very interesting concession to SAP’s strategy that Oracle are following suit.

Does SAP have first mover advantage?

In the analytics game – not really. SAP HANA and Oracle Exalytics are likely to be roughly as mature as each other and Oracle has a simpler sales-cycle because it is a one-stop-shop. This means that the 6 months early mover advantage that SAP has is not that relevant in terms of analytics.

However that’s not the point, because SAP HANA may be an analytics appliance today in its SAP HANA 1.0 SP02 guise – but it becomes so much more when SAP release SAP HANA 1.0 SP03 to market, which is purported to be in early November 2011. When this happens, HANA is no longer just a bolt-on analytics appliance, but also a full RDBMS for SAP NetWeaver BW systems.

Conclusions

For my money, Oracle Exalytics is just the second move of a chess game. Black Pawn to 2.d5. What comes next is what will be really interesting. Because SAP HANA has a fantastic roadmap that could position it as the RDBMS of choice for Enterprise IT – for SAP’s Business Suite software but also for so much more.

The question is, what does Oracle have up its sleeve to turn its RDBMS strategy into something that can compete with SAP? A collection of technologies will not be enough in years to come – it will have to be put into a cohesive strategy for both on-premise and cloud solutions. Currently SAP appears to have the upper hand in terms of long-term strategy, but Oracle should not be underestimated.

Updates

I’ve been reading some more and it looks nebulously like Exalytics can do both failover and scale-out – but the details in the solution brief are nebulous. Oracle also talks about Apps, which suggests it is trying to compete with SAP, perhaps with a future range of Oracle Exalytics Apps.

It’s also worth noting that Exadata and Exalytics appliances use an Infiniband 40GB/sec connection to connect each other, which may resolve some of the problems SAP is having around performance of scale-out solutions.