What Makes SAP HANA tick?

The colossal amount of big data is becoming tougher to handle for organizations. What can help the organizations crunch data and help in real-time analytics at least one order of magnitude faster? CIOs today are beginning to rely more on what is called in-memory computing (IMC).

In-memory computing (IMC) has for some time now been letting user organizations develop applications that run advanced queries or perform complex transactions on very large data sets faster, and is far more scalable than when using conventional architectures.

IMC Gaining Mileage

Organizations are transitioning to this by storing application data in a computer's main DRAM, rather than on electromagnetic disks. The main cause of rapidly adopting IMC is the ongoing maturation of application infrastructure technologies and the cost of semiconductor technologies going southward, say analysts.

"The relentless declines in DRAM and NAND flash memory prices, the advent of solid-state drive technology and the maturation of specific software platforms have enabled IMC to become more affordable and impactful for IT organizations," says Massimo Pezzini, vice president and Gartner Fellow. IMC technology is now more affordable and more proven.

Competing IMC Applications

With more than 50 software vendors in the IMC application infrastructure market, one such platform and solutions is the SAP HANA which banks on in-memory-enabling application infrastructure technologies and supports in-memory database management systems (SAP HANA DB), etc.

"Growing popularity of cloud and price reductions on SSDs/RAM makers have led to the emergence of new breed of NoSQL and in-memory computing softwares in the last few years. SAP HANA is one of them," says Harish Ganesan, co-founder & CTO, 8KMiles.com and an expert in big data applications.

When Mark Hurd, co-president, Oracle recently took a dig at SAP's HANA being their "most innovative database - then good luck to them", it did not come as a surprise. Amongst competing in-memory databases for online transaction processing and analytics workloads, IBM solidDB-a DB2 front-ended high-speed database suite, Oracle's In-Memory Database Cache, a performance extension of Oracle 11g, etc, rank close with their innovations. Comparable to HANA tools for analytics is Oracle's Exalytics.

However, Rajamani Srinivasan, vice president, applications sales, SAP Indian Sub-continent chooses to ignore competition, "We currently do not have any ERP or business suite application vendor who offers their application on in-memory or a real-time database like HANA. Lot of them claim that they have in-memory computing, but the complexity and the various levels of maturity exists-they offer more as an analytics engine which is what SAP was doing in 2010, which was the 1st generation when SAP started, but SAP moved on to offer multiple things; one of them being to offer the transformational business suite to be certified on HANA which is not something anybody else offers at this point of time. Eg, Oracle application or Fusion, I have not heard them offering Exadata, for example. It can be as a side car for Oracle apps but, I have not heard, frankly."

HANA is making it a battle of it. It was the biggest revenue yielding product line and fastest growing product line for SAP globally. "Second year of HANA we did some $350 mn (US dollars) globally. It is the fastest growing segment for SAP," reveals Srinivasan.

Seeking Faster Analysis

A recent IDC global survey of over 750 IT managers and business managers on challenges they face with separate platforms for transactions and analytics, found 40% required more than 2 days to prepare financial data for reporting. Respondents chose ‘faster access to information' as the factor that would have the most positive impact on their organizations. ‘Faster analysis' was chosen as the third most impactful change.

One implication of HANA's ability to work with a full database in-memory for analytics is that KPI computations can be completed faster with improved query performance when compared to disk based databases. "One of the biggest advantages that I see in SAP HANA is the ability to aggregate and update large volumes of data in near-real-time ie, it can handle both transactional (OLTP) and analytical (OLAP) processing effectively. Even AWS has a version ported to its cloud infrastructure called SAP HANA One recently, which enables companies of any size to deploy business-critical or consumer-facing applications that can leverage the power of in-memory transactional and analytical data processing offered by SAP HANA," says Ganesan.

Better utilization of current applications is another. "One benefit is that adoption of BI has tremendously increased; we are having 75-80% utilization of HANA for BI purpose and most of the users are now showing productivity increase in terms of creating reports from this rather than creating like earlier using excel etc. Now they are doing it spontaneously, so the report which used to get generated in half an hour earlier is now getting created in a minute," says Subodh Dubey, CIO, Usha International, sharing his experience on implementing HANA. "Asian Paints is also working on a plan to see how they can integrate. For eg, their CRM has been integrated with social media. Today from Facebook they get lot of data into their CRM. They are looking at how they can put it on HANA and make it run," Srinivasan adds.

Other key features that keep it going are:

HANA has progressed to running on servers with 100 TB main memory powered by IBM by May 2012.

Compatibility with other database products. As a database agnostic company, it also partners with vendors such as IBM, Oracle and Microsoft for business applications. Its own offering, SAP Sybase SQL Anywhere, is also compatible with HANA. "SAP is advancing its real-time database capabilities by integrating key Sybase database technologies such as SAP Sybase SQL Anywhere into its comprehensive enterprise information platform," Carl Olofson, research vice president for application development and deployment, IDC stated recently. "The new version of SAP Sybase SQL Anywhere technologies work in combination with solutions such as SAP HANA decreasing administrative complexity, allowing SAP to enable practical, easy to use enterprise data delivery to users anywhere, any time," adds Olofson.

"From a pure database management standpoint, our entire business suite landscape has been certified on HANA. HANA can be implemented as a standalone for advanced analytics capabilities which can be done in a non business suite environment," says Srinivasan.

What about Glitches?

One of the important questions people ask regarding in-memory computing systems is what happens when the server (power) fails? "SAP HANA supports full HA scenarios and standby nodes in case of hardware failures. SAP HANA writes a copy of what is happening in memory to disk, using a combination of save-points and log files. So you can DR using disk mirroring to an alternative data center as well," says Ganesan.

However, there is another challenge too in terms of adapting design changes while adopting HANA. "One challenge I can observe is that, because SAP HANA simplifies the way how applications can be architected, it sometimes requires a radical change in the design philosophy for the enterprise architects," Ganesan adds. "Yes, we had to go in for design changes for some of the reports as well because they were not as per the standard for HANA and these reports we had to make the design change and that took about 2 months time. We are still doing the design changes, it is a continuous process and how to create new and analytics based reports, etc," reiterates Subodh Dubey, CIO, Usha International.

With the clock ticking since launch, HANA is thus beginning to impact analytics, benefiting organizations - if not real-time, at least in near-real-time.