SAP and SAS partner for fast big data in-memory analytics

Madan Sheina, Lead Analyst, Software – Information Management

SAP and SAS announced a new partnership to create a joint product roadmap for in-memory analytics solutions. The aim is to make SAS’s advanced analytics software run smoothly on SAP’s HANA in-memory platform, helping joint customers run bigger, more complex, and faster analysis workloads.

The combination of technologies is potentially compelling for customers looking to mine intelligence out of Big Data. However, depending on which side’s perspective you take, this is a curious tie-up between two fierce analytic rivals that each market and sell their own in-memory platforms.

It remains to be seen whether this partnership can bring out the best of SAP and SAS’s core analytic strengths and competencies so that both parties gain, while also creating compelling value for customers. Historically, this hasn’t always been the case. Nevertheless this is one of the more interesting high-profile partnerships in analytics to watch in 2014.


Co-innovation points to bilateral development and product

The technical specifics of the partnership relate to joint product engineering and integration and co-marketing and selling of solutions, which will be prioritized on two levels: horizontal analytic processes requiring advanced analytics to run on an in-memory platform (e.g. customer and marketing intelligence, risk and asset management, anti-fraud/money laundering); and industry verticals (initially financial services, telco, retail, consumer products, and manufacturing).

Neither company has thought too deeply as yet about what a joint product will look like. Ovum believes it will take a few months for each company to figure out where its technology fits into a joint solution offering.

The stated joint product roadmap aims to have SAS’s advanced analytic algorithms run directly in the HANA database – i.e. pushing analytic processing to where the data sits to improve performance. A logical architectural alignment assumes that SAP HANA will be the in-database analytic platform in common accounts, leaving SAS’s server-based analytics to work out where the statistics and mathematical calculations processing is performed. Some integration is already available. The SAS/Access module already provides a realtime read-write interface to HANA.

Both parties bring something to the table. SAS has been developing extreme predictive analytics for many years, and has more PhD data scientists on staff than most vendors and enterprises. SAP, meanwhile, has focused largely on more mainstream business analytics users, touting parallelization and scale-out systems.

If SAP and SAS get the integration right, joint customers will benefit from a simplified Big Data analytics architecture that minimizes data movement, duplication, and reconciliation between the two environments, enabling massive parallelization of computationally-intense workloads, all in-memory and in-database.

A toe in the water with customers and partners

This partnership is still in an exploratory stage. The first step will be to enlist joint customers (which neither has announced as yet) for a co-sell pilot program to validate SAS analytics running on HANA. Both companies also assert their prime global consulting firms and system integrators are driving forces for the partnership – notably Deloitte and Accenture. These partners thrive on opportunities to assemble custom analytic systems using best-of-breed technologies. One sticking point, however, could be incompatible licensing costs – something that SAP and SAS will have to iron out and test with customers.

Perspective from both sides is necessary to gauge success

For SAS, this partnership builds on its long-term strategy of improving the performance of its offerings. HANA gives SAS a high-speed processing platform option. One of the issues that SAS has struggled with when running on an aging architecture is latency. Running advanced predictive analytics in SAS often needs to transfer data from storage to its own analytic server engines, which takes time, especially with large volumes of data. It also promises to open up new sales opportunities for SAS by penetrating into SAP’s vast installed base. For SAP it will provide more impetus for HANA uptake, which the company banks on as a key growth factor for 2014, and will extend the database’s range of use cases to more advanced forms of predictive analytics.

While the technologies potentially align, success will ultimately depend on how intelligently the companies work together for common benefit rather than trying to score marketing brownie points over one another. Consider that both companies have in-memory platforms aimed at analytics.

Last year, SAS announced its LASR Analytic Server, which performs in-memory analytics stored on Hadoop. And it was not so long ago that SAS seemingly took aim at SAP with in-memory analytics, announcing its own server-blade-based in-memory capabilities as part of several new high-performance analytic products, taking cheeky potshots at HANA as a SQL-bound database. SAP also recently took a shot across SAS’s bow when it acquired KXEN, a developer of predictive analytics software, which is of course a stronghold for SAS.

It remains to be seen how this partnership will play out. SAP might be more bullish on this one compared to others since it seems to be a much deeper technology play and is focused on common customers. From SAS’s perspective, much will depend on whether any group within the company sees HANA as a direct threat. Regardless of whether both parties’ engineering teams forge a cordial relationship, the key question is whether their marketing departments behave.

Signals a new era of “co-opetition” among analytics vendors

SAS and SAP will remain competitors even after this partnership. SAP will continue to push its BusinessObjects BI portfolio and its KXEN-based predictive solutions. SAS will undoubtedly do the same with its competing BI and analytics products, but will also ensure its software works well with HANA and will continue to partner with other vendors like Teradata for in-database analytics. Since companies increasingly face a complex future, where multiple analytic tools and multiple data platforms are now an IT reality, this kind of co-opetition is perhaps welcome, especially for joint SAP and SAS customers that want to run Big Data applications and workloads. However, for joint customers seeking advanced in-memory analytics, what is the route to take: SAS LASR, SAP HANA, SAS Analytics, or SAP KXEN?

Related Stories

Leave a comment

Alternatively

This will only be used to quickly provide signup information and will not allow us to post to your account or appear on your timeline.