SAPPHIRE is just two weeks away and a big theme at this year’s conference will undoubtedly be in-memory computing. And the stakes are high for SAP.

Data analytics today is all about size and speed. The two key questions vendors need to answer are 1) How much data can you process? and 2) How fast can you do it? SAP is counting on in-memory computing, in the form of its SAP High-Performance Analytic Appliance, or HANA, to answer both questions.

Released late last year, SAP HANA is a preconfigured database appliance that relies on in-memory storage to process and analyze large volumes of data in near-real time. SAP is in the process of rolling out a number of applications that sit on top of HANA, among them BusinessObjects Strategic Workforce Planning.

Here’s what makes in-memory computing important. In traditional business intelligence (BI) and data warehousing environments, data is modeled and then loaded onto disk, a process that can take weeks or longer depending on the amount of data and complexity of the models. Users run analytics against the data residing on disk. With in-memory technology, data modeling is no longer necessary as data is loaded into random access memory instead of disk. Users run queries against the data inside the application or database itself.

Bottom-line: In-memory computing adds speed and scale to traditional BI applications.

But an in-memory appliance like HANA is only as useful as the applications that run on top of it, so there’s a lot riding on SAP’s application development pipeline. The first HANA application out of the gate, the aforementioned BusinessObjects Strategic Workforce Planning, was a dud.

The application lets companies model the impact of mergers, acquisitions or other events on staffing levels. That’s a fine application, especially useful to companies in volatile industries with a lot of M&A activity, but it doesn’t require in-memory computing to back it up. It’s not a data heavy application. There’s only so much HR data at any given company, compared to, say, social media data or sensor-generated data. Nor does HR analytics need to be real-time.

So it doesn’t make sense for a company to spend extra money on an in-memory appliance like HANA in order to run HR analytics. You can just run an application from Taleo or Workscape with batch-loaded data from your HR database. It’ll cost you a lot less and get you pretty much the same results.

Among the other HANA applications SAP is reportedly developing is a smart meter analytics app. That, to me, is the type of application tailor made for in-memory computing: lots of data that needs to be processed and analyzed quickly so adjustments can be made to power flow in near real-time.

It’ll be interesting to see the next HANA apps SAP rolls out, which may happen at SAPPHIRE in a coupe of weeks. For SAP’s sake, hopefully they’ll be more in line with the smart meter app than the HR analytics app. Otherwise, it will have invested a lot of money developing an in-memory appliance that nobody uses.