SAS CEO Jim Goodnight: Not The Retiring Type
Cloud, shmoud! Hadoop, big whoop! SAS' top executive and co-founder throws cold water on hot trends. Behind the times, or a clear-headed realist?
This year has seen some notable CEO departures from the tech industry. Some moves have been predictable, like Sam Palmisano stepping down as IBM CEO at the end of this year to become the company's chairman. Some have been lamentable, like Steve Jobs' passing. And some have been inglorious, like the close of Leo Apotheker's brief tenure at Hewlett-Packard.
And then there's Jim Goodnight. Co-founder and CEO of SAS Institute, Goodnight, at 68, is as steadfast and singularly focused on SAS's success as ever. And as always, Goodnight doesn't mince words, hew to conventional wisdom, or worry about being perceived as old school.
More Software Insights
- The Critical Importance of High Performance Data Integration for Big Data Analytics
- Mobile DevOps: Achieving continuous delivery with multiple front ends and complex backends in Banking, Financial Services, and Insurance
- Managing Volatility through Smart Inventory Planning
- Agile Maturity Report: Benchmarks and guidelines to improve your effectiveness
Cloud computing? "I think it's a lot of hype," he says during a recent interview in the company's midtown Manhattan office. "Time sharing made a lot of sense when computers cost $3 million in 1970, but now they're commodity items that anybody can buy."
He points to the power and affordability of Intel's Westmere chip, detailing the number of cores, threads, and parallel processes it can run simultaneously.
[ Want more on analytics competitors? Read Low-Cost Options For Predictive Analytics Challenge SAS, IBM . ]
Yes, but with the likes of IBM, Oracle and Microsoft joining Amazon in the public cloud game, won't that kind of processing capacity be available cheaply and on demand? "You can buy one of these servers for $500 and have it under your desk," Goodnight responds.
Hosted analytic applications are a fast-growing part of SAS's business, expected to surpass $100 million out of more than $2 billion in revenue this year. But this is conventional, on-premises software run by SAS, with relatively fixed licensing terms and no provisions for cloud-like elastic scaling. Goodnight doesn't dwell on this model. He allows that private cloud and virtualization approaches are here to stay, with good reason. "But for general-purpose computing, the issue of letting corporate data be out on some kind of cloud where you don't have any idea where it resides is not something a lot of companies are going to embrace," he says.
We talk at length about SAS's in-database and in-memory processing capabilities and he explains "the really exciting work" the company is doing on SAS High-Performance Analytics, software due in December that will speed data modeling with the power of massively parallel processing on appliances from partners EMC and Teradata. "A risk-analysis job that used to take 18 hours can now be processed in about 12 minutes," he says.
When I bring up Hadoop, the open-source data processing platform that hundreds of large-scale Web 2.0 power players have embraced, Goodnight says SAS has a connector to HDFS (the Hadoop Distributed File System), but he's clearly not worked up about it. "Hadoop is still open source," he says, as if that's a dirty word. "It's not a completely cooked thing to be investing millions and millions of dollars in."
Isn't it telling that EMC, IBM, Microsoft and Oracle have all embraced Hadoop with their own distributions or plans for distributions announced this year? "That's probably a case of, 'I don’t want to be left out, so I'll make an announcement, too,'" he says. "If companies are putting their data in Hadoop, that's fine with us, that's where we'll read it. If it's in Oracle, that's where we'll read it. If it's in DB2, that's where we'll read it."
Goodnight's not being quite as clear as he could be, since competitors such as Revolution Analytics with analytics and even mature vendors like Informatica with its parsing code are bringing their software inside Hadoop. "That's where we'll read it" would seem to mean copying or moving big data into a conventional relational environment, which wastes time. SAS has partnered with the likes of EMC and Teradata to apply the analytics where the data lives, and to take advantage of the processing power of a scalable platform. But SAS has yet to announce any such capabilities where Hadoop is concerned.