Definitely. We see for the first time very significant market opportunity for SAP in this area. We’ve really not been into this market aggressively. We have some of the Sybase technology, and now the HANA technology. That has opened up a significant market opportunity we didn’t have before.
Look-Back Interview: Will SAP's HANA Transform Financial Management?
When SAP's in-memory HANA machine was introduced in 2011, Computerworld UK said it could cause the biggest paradigm shift in enterprise computing since the introduction of IBM's original PC in the 1980s.
“HANA can hold all your finance data, all your risk data, and . . . it will also hold your software programmes,” says Andy Hirst, Vice President for Banking Solutions at the German enterprise software giant. “So we’ll have the ERP in there as well.” Because the data and software are embedded in-memory rather than in hard disks, access and processing are much faster – 100,000 times improved query performance, claims SAP.
Hirst spoke to CFO Innovation’s Cesar Bacani at the Sibos conference in Osaka in November 2012. Excerpts from this look-back interview:
Is SAP itself using HANA machines?
Absolutely. For sales forecasting, for finance operations, we’re featuring our own products, GL and ERP, and we run them already for all our sales on demand, for management pipeline, for our positioning capabilities.
What we found is, when you start to have real time access to information, you can build a number of front-end apps that allow you to look at gazillions of pieces of information [through mobile devices].
So you may have all your liquidity risk information for a bank in one HANA machine, and then you would use an online app like this [on my smartphone] to see your current capital risk ratio, what your stable funding ratio is in the future.
You might want to drill down and see how you are doing against certain plans. You might then want to perhaps look at different scenarios, you might want to look at certain groups and subgroups, you might want to redo that calculation for a billion records and run it straight from here [your smartphone] and it will reformat in front of your eyes.
How stable is HANA? Aren’t there some glitches because it processes so fast?
Like every database technology, we have had to put all these common things, like somebody comes along and pulls the power cord out, failover, in the event of a storm, would you have industry recovery solutions . . . All those have been thought through.
We’ve got many challenges on a daily basis, but we actually cope with them . . . . With each release, each year we’ll get more and more mature in managing the fault-tolerant nature of HANA machines. We will catch most of those in the next few months, but regenerating [data], resetting after corruption, all those are currently possible.
We have many solid individuals who came from Sybase who’s got 28 years of database knowledge. [Editor’s Note: SAP bought Sybase for US$5.8 billion in 2010].
If you invest in HANA, does it mean you can no longer use your existing servers?
The great thing about HANA is that it is non-disruptive. It allows you to extract information from any system, any mainframe system, any server system, any other competitive third-party application, whether you bought it for yourself or bought it for somebody else.
It extracts the data, holds the data, allows you to do all the analysis. Once you have post ERP or post transaction banking, all the information is locked in those systems. You get to unlock it, put into an in-memory environment, and then to unlock that data to make better business decisions, more general wisdom or piece of information to help somebody make a better decisions at the frontline.
You store the data in-memory. You don’t need [conventional] storage anymore. These [USB] devices changed PCs because they can store in flash memory. We’re doing this in an industrial scale. HANA can store 100 terabytes, which is just a mind-boggling amount of information.
And soon it will be able to host ERP software, too.
Next year we are going to be able to [host] the transactions ERP. This will transform the architecture of the future for the finance director. When we can tell our 2,000 ERP customers, you will now have the option to hold all that information which you [currently] hold in disk drives, which took time to respond to do the closing of the books, you can hold all that in-memory, this is transformational. It dramatically cuts the cost of setting up the finance operation and getting access to the information.
You’ve got to remember, a lot of the world’s trade commerce works on SAP ERP, finance and GLs [through SAP customers in] utilities, oil and gas, and banking as well. We have six of the 12 top banks in the world using our GL. We have many planning, budgeting consolidation customers, many reporting customers. So this is a story that we see, a very strong vision, going forward.
It doesn’t matter what business intelligence software you use?
What [customers] tend to do is they use Business Objects, which is SAP’s. We make it very attractive; it integrates well. But HANA is an open platform. It will work, for instance, with Panopticon, if you want a very strong real-time visualization layer.
If Cognos [a business intelligence software company bought by IBM in 2007] wants to port, they could, but they have chosen not to. That’s their choice, but it’s an open platform . . . I think perhaps they think they might want to offer a competitive offering. So they might have their own DB2, or their own database technology that they would like to wrap with it.
But Cognos is just a different business intelligence tool that can access the HANA machine using SQL, if they choose to. They have not chosen to.
IBM has something similar to HANA going on, then?
Nobody has something similar. We have a market lead in that perspective, but you could say that [the competition has] various variants. None of them are truly in-memory. They have a cache, they have flash disks, but they do not have a truly large industrial scale in-memory area. And that’s the transformational part of the story for us.
How does HANA relate to cloud computing? You wouldn’t need cloud if you have HANA?
You could have both. You could, if you have an application that does not involve the most sensitive data, it’s usually not transactional data, for instance, you could hold all the information in the cloud, in a HANA machine, and just use it as a service. Amazon, for instance, has a service where you can use HANA in the cloud. You just upload all your data into the cloud, you buy so much space in that particular service and you pay a monthly fee.
That’s an option and we’re going to see that quickly adopted especially by midsize companies, which don’t want to invest in the whole box. They can buy a portion of the machine a portion of the time.
Why, is HANA hugely expensive?
Not really. I don’t have the exact pricing for you, but it ranges from the half-a-million dollar investment for a large corporate at the lower end, to some corporates now spending north of US$5 to US$10 million. But we also offer it to small and midsize companies at very attractive price points, 50k-100k.
You basically buy as a bundle. It has a database – you have certain tools such as predictive capabilities that are already bundled in the machine. All you need to buy is the BI or the ability to access the information through a license of a business intelligence tool, either ours or Panopticon or some other provider.
SAP is a software company, though. Do you make the HANA machines, too?
You can buy the hardware from IBM, from HP, from Cisco, from Fujitsu, HP, Hitachi, NEC – they all make it specifically for SAP . . . They build it to spec, and they have slightly different price points. But they all pre-load the software, and the database and all the HANA-SAP content.
Do you see current data centres being replaced by HANA data centres?
Read more on