Hasso Plattner, who 20 years ago designed a computer program that supercharged SAP AG's growth, has been pursuing another breakthrough that could determine the software giant's fate.
Now SAP's chairman, the 68-year-old engineer is trying to take advantage of cheaper memory chips in servers to speed up complex business calculations and allow companies to do in seconds what currently can take hours or days. The aim is to allow executives to quickly access and analyze business data even on hand-held devices.
If Mr. Plattner succeeds, he hopes to revolutionize business computing again and put his main competitor, Oracle Corp., on the defensive. But if he fails, SAP could end up stagnating in an industry full of bigger and richer tech adversaries.
Oracle Chief Executive Larry Ellison publicly derided Mr. Plattner's big bet as "whacko" in 2010 and said he wanted the name of SAP's "pharmacist." Counters Mr. Plattner: "If competitors are joking about you, they are vulnerable."
For his bet, Mr. Plattner decided to do an end run around SAP's corporate research-and-development department with thousands of engineers.
Instead, he recruited a bunch of university students in this small city outside Berlin. Working in a converted East German railway building dubbed "the villa," these T-shirt-clad 20-somethings built the prototype of Mr. Plattner's new product.
"It's not so easy to break out as a large company and do something radically different," says Mr. Plattner. "At the university, you have the freedom."
What the students came up with—known initially as Hasso's New Architecture, and now called HANA—allows companies to store data in servers' main memory, instead of using the relational databases that Oracle dominates.
During its fourth-quarter news conference Wednesday, SAP Co-Chief Executive Officer Jim Hagemann Snabe called HANA "probably the biggest innovation in the business software industry in the last 20 years." He said it would allow HANA to grab leadership in the database business and in cloud computing—selling software as a metered service to customers over the Internet. SAP said HANA reached €160 million, or about $208 million, in sales in 2011, ahead of its earlier goal of €100 million.
For SAP's Mr. Plattner, the challenge isn't only about business. For years, the German executive has competed with Oracle's Mr. Ellison, a fellow yacht racer, in both software and sailing.
In Oracle's second-quarter conference call, Mr. Ellison said that the company had won one deal against SAP's HANA using its Exadata database machine. He added that an Oracle customer, which he didn't name, had tested the Exadata database machine against SAP's HANA and found it was faster than HANA. "That came as a surprise to us," he said.
Mr. Snabe Wednesday dismissed that claim, saying a comparison needed to be conducted by a third party with transparent methodology to have credibility.
SAP already is the biggest player in software that businesses like retailers and manufacturers use to track sales and inventory. Its eventual goal with HANA is to sell software that tracks transactions and performs data analysis, and does it all it in seconds. It expects to gradually evolve the software over the next two to three years, though some analysts think it could take longer.
There is a reason SAP is aiming high. The company's core business of selling business applications, which large companies use to track everything from sales and inventory to customer contacts, is under attack by Oracle. SAP's share of the $58 billion business-application software fell to 14.8% in 2010 from 15.5% in 2007, while Oracle's jumped to 10.7% from 9.9% in the same period.
Mr. Plattner has pledged to rework all of SAP's software so it can run on HANA. "They're betting their business on it." says Yefim Natis, an analyst with Gartner Inc.
In-memory computing could be crucial for cloud computing, because offering services online requires companies to rapidly process large volumes of data. In December, SAP said it would pay $3.4 billion to acquire San Mateo, Calif.-based SuccessFactors Inc., which offers online services that help manage employees and carry out performance reviews. The company also paid $5.8 billion in 2010 to acquire Sybase Inc., which makes software that can send business information securely to mobile workers on their devices, easing a potential concern with HANA.
SAP started a HANA pilot program in 2010, offering the software as part of an integrated system to Procter & Gamble Co., Nestle SA and others, before selling it generally in June.
SAP has pledged to grow past €20 billion in revenue by 2015, on the back of HANA. It said Wednesday that it posted €14.23 billion in revenue for 2011, up 14%. SAP's net profit more than doubled in 2011 to €1.2 billion from €434 million a year earlier.
SAP convinced Charité Universitätsmedizin Berlin, a large university hospital, to drop its Oracle software and switch to HANA. Together SAP and Charité developed a prototype of an iPad software application that uses the HANA machine to analyze three million data points for 140,000 admitted patients annually and determine if they are a fit for a clinical trial. Using the application, the hospital said it reduced the time it takes to find patients from weeks to less than one second.
"We need all this data in real time," says Martin Peuker, deputy chief information officer for the hospital.
Carrying out Mr. Plattner's grand plan won't be easy. For one, rewriting its software, which has roughly 400 million lines of code, will take years. SAP must convince companies to change technologies and practices they have held for a long time. Then there is Oracle, which in October launched a new in-memory analytics system aimed directly at SAP.
Mr. Plattner is one of five German engineers who started SAP in 1972. In the early 1990s, he anticipated corporate computing's shift from bulky mainframes to networks of personal computers and created business-application software for this new environment. He was SAP's largest individual shareholder with a 9.96% stake worth more than $6 billion as of December 2010.
In 1998, Mr. Plattner founded the Hasso Plattner Institute as a software-engineering affiliate of Potsdam University, and began teaching there. "I couldn't impress them," says Mr. Plattner of the young people in his class, who were besotted with Facebook Inc. and Google Inc. but saw SAP as stodgy.
So, in 2006, Mr. Plattner decided to shake things up. He took a bottle of red wine and sheet of paper into the garden behind his house. By the time he had reached the bottom of the bottle, he says, there wasn't much written on the paper. But, he says, he reached the conclusion that in-memory systems were the future.
Soon after, Mr. Plattner formed a group of three doctoral students and a handful of undergraduates at the institute to study in-memory technology. He sketched a diagram of a new database model on a whiteboard and told them to start building.
"Hasso has a lot of presence," says Jan Schaffner, one of the Ph.D. students. "If he's talking and you don't understand something, you have to bring some guts together to interrupt him."
The students started experimenting with existing SAP database software as they worked to develop a new prototype. Mr. Plattner arranged for SAP software experts, including Vishal Sikka, now the company's chief technology officer, to teach them.
The students toiled away in the institute, their work space fitted out with red couches, a large screen TV and a foosball table hooked up to an iPad that serves as a digital scoreboard. By June 2007, they were ready to unveil their first prototype. They traveled to SAP's headquarters in Walldorf to show it to Mr. Plattner.
In an auditorium at the company's sprawling glass-and-steel complex, the students found an audience of dozens of SAP's top engineers. As the students took turns nervously explaining aspects of the project, some of the SAP developers interrupted to point out problems with the prototype. One concern was what happens to the data stored in memory if the power goes out.
Mr. Plattner jumped up to defend their work. "Don't you get it? It's not about a product now," he said.
"He was really passionate about it," says Christian Schwarz, 27, who at the time was an undergraduate assisting with the presentation.
Mr. Plattner used his contacts to advance the project. He met former Colgate-Palmolive Co. Chief Information Officer Ed Toben in the Orlando, Fla., airport, and persuaded him to provide transaction data for the project.
He appealed to former Intel chairman Craig Barrett to secure the latest microprocessors for the institute's labs. To be closer to the student team, Mr. Plattner bought a nearby house. Winston Churchill stayed in the house when he attended the Potsdam Conference in 1945 that helped determine the future of Europe after World War II.
Meanwhile, Mr. Plattner's quest had become more urgent. After the global financial crisis hit in 2008, companies pulled back sharply on tech spending. Under former co-chief executive Léo Apotheker, SAP raised its maintenance fees in an effort to make up for lost licensing revenue.
Customers revolted against the service price hike, and SAP had to rescind it. Profits started to erode in 2009. Mr. Apotheker stepped down. Mr. Plattner and the board promoted Mr. Snabe and Bill McDermott to be co-CEOs. Mr. Plattner also took on an expanded role as chief software adviser.
Mr. Plattner takes on a new group of six to eight undergraduate students a year to conduct research projects. By the summer of 2009, a new group of students had a working database prototype that could filter through 250 million customer records, find 380,000 unpaid invoices and then single out the roughly 200,000 overdue bills in just 1.5 seconds. That compares with the roughly 20 minutes it would take standard software to do this, the students say.
Once again Mr. Plattner had the students demonstrate the prototype to a group of SAP executives in Walldorf. Some in the room were skeptical. But Mr. Plattner told them that such speed was necessary for SAP's survival.
Inside SAP, management allowed Mr. Plattner to create a new database group to turn the prototype into its next product. Mr. Plattner invited his students to join in and make suggestions for improvements.
SAP approached Erez Yarkoni, chief information officer of T-Mobile USA Inc., the U.S. unit of German operator Deutsche Telekom AG, last April, to pitch the new wares.
Mr. Yarkoni says he was skeptical at first because a decade earlier when he worked for another company, he implemented an in-memory database from a company called TimesTen, which Oracle later bought. TimesTen promised the database would give him 10 times the performance. It didn't, he says. Oracle declined to comment.
SAP promised Mr. Yarkoni that if HANA didn't work, the company wouldn't have to pay.
T-Mobile wants to lure its customers to upgrade its data plans. To do that, the U.S. operator targets 33 million customers with almost countless variations of offers for different smartphones or data plans.
The goal is to see what works, and repeat, Mr. Yarkoni says. Once someone responds to an offer, the answer comes in through software running on servers in T-Mobile's data center. From there it heads into a data warehouse, where it is stored on a hard disk.
The problem, Mr. Yarkoni says, is that it took a week to three days to find out seemingly simple things such as who signed up for what. The bottleneck, he says, was the hard-disk storage, which uses rotating parts and a motor to access the data stored on it.
Working with SAP, T-Mobile installed HANA in October. Instead of feeding the results into a disk-based data warehouse, it went into the memory of a HANA database machine. It spit out an answer in 15 minutes. "If it takes you a week to understand and refine the offer, you probably are missing the sweet spot with the customer," Mr. Yarkoni says.
read more: Olympus Wealth Management
No comments:
Post a Comment