Atul Patel, Vice-President SAP Analytics, SAP Asia Pacific Japan, talked with Jasmine Desai about SAP’s renewed investment in the analytics space with a focus on social, collaborative and mobile analytics along with its strategy for in-memory
SAP has a strategy around five market categories namely applications, mobility, analytics, BI and database technology. Enterprise analytics and applications work hand in hand. The fourth version of our analytics platform was released in 2011. The analytics suite works with SAP and any other software from a transactional viewpoint. If one wants to get analytics into the hands of consumers, however, one needs to have different types of consumption models, for which we need to go beyond the desktop and mobile-enable it. That’s why we invested in Sybase. It has a mobile development platform through which we can make applications available irrespective of the platform or device. From the deployment standpoint, the Cloud is becoming important. With the acquisition of SuccessFactors, we invested heavily into the Cloud. Historically, data has been kept in databases. The relational database concept had not changed for twenty years. In the past two years, we have invested heavily on in-memory computing technology with HANA that makes it possible to have OLTP and transactional data both coming out of the same database without any performance penalty.
In-memory is believed to be an expensive option for the Indian market. Do you see it gaining widespread adoption?
Organizations have relational databases, operational data store, data marts, data warehouses, analytical warehouses etc. There are lots of layers. Part of that can be blamed on the IT industry that has convinced organizations that they need to have so many islands of data marts or data warehouses. HANA enables you to get rid of all of these layers. The price of memory has come down. There’s no adverse impact on performance because of doing column-based indexing in memory. You can have analytics in the same database as well. There is a lot of benefit to having real time analytics from the existing environment.
The cost of managing different data environments is effectively gone with a HANA implementation. HANA can handle both transactional and analytical work in same database. It has been written from the ground up to have a columnar structure and also to take advantage of in-memory. The gain that one gets by running HANA in-memory is more than putting a traditional database in-memory. The price is still high in India but, when you remove the layers, there are huge savings in terms of managing operations.
People say that HANA works better with structured data. What’s your take on that?
SAP has Enterprise Information Management capabilities that it got with the acquisition of BusinessObjects. It has the capability to look at unstructured data and make sense of that. This is integrated with HANA. Hadoop and HANA are complementary.
If you are doing some work with respect to big data in Hadoop it can be combined with HANA to create a ‘universe’. One of our customers, Nomura Research in Japan, uses HANA for unstructured data analysis for genome cancer research.
Most organizations in India are on Netezza or Exadata? What is SAP’s strategy to make inroads in this market?
Netezza does not have OLTP plus OLAP. It is just an analytical appliance. Customers do not have to move right now but, if they are looking at a three to five year horizon, and if the business is growing, then they will consider us. We have taken advantage of Budgeting, Planning and Consolidation (BPC) on top of HANA. It has BI, EPA, GRC and collaboration.
We now have capability to do in-memory data mining. A lot of data crunching happens when doing statistical mining. We run predictive analytics, which is the front end on HANA. The number of records for which mining can be done is unlimited in HANA.
Could you please elaborate on your SI relationships in this sphere?
When it comes to analytics, partners are key. 40% of our business in the APJ will come from partners this year. HP, IBM, Fujitsu and Cisco are all strategic partners for HANA. On the analytics side, we have SIs and resellers. E.g. Oracle’s iFlex has BusinessObjects embedded in the solution.
What do the entry level and mid-market points look like in India?
We have made HANA available for all segments. We have HANA Business One, HANA for BW and HANA Edge for SMBs. With respective to predictive analytics, it is based on the number of users. For analytics, there is a huge uptake in retail. There is also a lot of strategic investment happening in the government.