The CEO Perspective
Mike Koehler and Jim Goodnight on the SAS, Teradata Partnership
Five months after unveiling its strategic partnership with Teradata, SAS announced at SAS Global Forum that the two partners had reached significant milestones, allowing joint customers to benefit from in-database analytics sooner than anticipated. In addition, the two companies have launched a new program to prepackage select SAS software on Teradata-managed servers within Teradata environments.
Together, SAS and Teradata are making it easier for customers to exploit SAS Analytics through the core parallel processing functionality inherent in Teradata’s architecture. The benefits? Faster predictive results, improved accuracy of analysis, reduced data movement and streamlined analytic processes.
To learn more about the strategic aspects of partnership and how it will continue to benefit customers, we garnered these insights from the leaders of both companies: Mike Koehler of Teradata and Jim Goodnight of SAS.
How does the SAS and Teradata partnership help mutual customers address their current business challenges?
Mike Koehler: Businesses face a host of threats and opportunities – competition, globalization, regulations and economic pressures. The companies that can see and act on trends before the competition does, win. This partnership helps organizations optimize how they process and use data throughout their businesses. Ultimately, that gives them the intelligence and speed they need to make the best decisions and take actions faster than the competition.
Jim Goodnight: Our mutual customers, I think, are all very happy about this move because, up until now, they would store large amounts of data in their Teradata system, and then they would have to do an extraction to bring it over to SAS so that SAS could then analyze the data. With this effort to move more of SAS inside the database, that’s going to speed up processing by leaps and bounds.
Tell us about the areas where SAS and Teradata will work together to deliver value to the marketplace.
Koehler: The partnership with SAS is both strategic and comprehensive, so it covers engineering optimization, joint solutions, technology road maps, and joint selling and marketing. Our goal is to optimize SAS analytic solutions for the Teradata Database.
Goodnight: One of the first areas that we’ve been working on is to deliver the new SAS Scoring Accelerator for Teradata, which scores models directly inside of the Teradata Database [see sidebar]. That will make it possible for huge amounts of data to be scored very, very rapidly.
Koehler: Another area that we’re already working on together is the Teradata and SAS Center of Excellence [CoE], which is a joint team of solution architects and technical consultants who are dedicated to helping customers optimize their SAS and Teradata environments.
Mike, what value do you think SAS brings to the partnership?
Koehler: The partnership gives our joint customers improved performance, reduced data movement, and faster and broader access to more analytic insights. For example, SAS brings a robust set of financial applications for credit risk, anti-money laundering, and fraud to the table – offering more choices to our customers to improve their enterprise-level decision making.
Jim, what value do you think Teradata brings to the partnership?
Goodnight: Teradata brings to the partnership vast experience in dealing with huge amounts of data over a grid network. We’re seeing more and more chips come out these days with multiple cores, which requires an extensive amount of reprogramming if you’re going to really take advantage of the multithreaded cores. Whereas with the multigrid chips, there’s only one I/O [input/output] channel to handle all the cores. This means as the amount of data continues to grow and grow over the years that it’s going be important to be able to move as much of the computational effort as close to the data as possible. Teradata gives us a platform to do that.
Describe some of the early accomplishments that are allowing joint customers to benefit from this partnership.
Koehler: We’re making good progress with this relationship, and one example is the new SAS Scoring Accelerator for Teradata that Jim mentioned. By leveraging scoring in the Teradata Database engine, customers can significantly increase processing speed and performance, and process more analytic models faster. These improvements will help customers use the results of analytics more quickly, allowing more time to execute plans and programs.
Can you tell us about the results you’ve seen in the initial benchmarks of the SAS Scoring Accelerator for Teradata?
Goodnight: An initial benchmark running on a Teradata system achieved a 4,500 percent performance increase compared with traditional SAS scoring. Traditional SAS server deployment processed 16,400 rows per second, where in-database scoring processed 125,000 rows per second on each Teradata node, resulting in 750,000 rows per second on a six-node Teradata system. The benchmark’s performance indicates there is linear scalability for SAS scoring in relation to the number of Teradata nodes deployed. An improvement of this order of magnitude shows how customers can significantly increase the number of analytical models they can simultaneously
process and deploy, accelerating time to value.
SAS has many partners; Teradata, I’m sure, has many partners. Tell me what’s unique about this particular partnership between the two companies?
Goodnight: The uniqueness of the partnership between Teradata and SAS is the fact that Teradata is the only vendor that we are working with right now to move SAS computation directly inside the database. This will give incredible additional speed and performance and also make jobs that used to run many hours be reduced to a matter of minutes.
Koehler: I agree, we’re uniquely offering the market a higher level of integration between our two solution sets. We’re also aggressively aligning our teams – at all levels of the organization – to enable customers to effectively compete on analytics: CEO commitment, a joint engineering road map, teaming of our field sales organizations, extensive co-marketing of our relationship, and dedicated engineers from both companies working together in the Teradata and SAS CoE.
Looking forward, is there anything that SAS and Teradata customers should know about where the two companies may go next?
Koehler: SAS and Teradata will continue to achieve key milestones to extend SAS Analytics and data transformations inside the Teradata Database. Our joint development plan also calls for SAS solutions to be optimized with the Teradata Database in key industries, such as financial services and retail.
Goodnight: Also, we’re putting together a council of customers right now that can sit on an advisory board for us to help make sure that we understand what the road map should look like for further integration. We both have some ideas of what things we need to go after right away. But, it’s really going to require customer input to help us guide our direction and help us develop our road map.
Will customers get a say in the solutions that will help them solve the business problems they face?
Goodnight: We need customers’ input considerably here because the amount of effort it would take to move the entire SAS system into Teradata would probably take four or five years. It’s because SAS is a very large system. And we want to do it in such a way that we can satisfy as many customers as possible as quickly as possible, so that’s going to require customer input.
Koehler: We constantly listen to our customers describing their priorities and challenges, and this influences how we invest and what we build. We have formal and informal vehicles to collect this feedback, such as our customer advisory boards, our direct interaction with customers, and surveys. As Jim mentioned, we have launched a SAS and Teradata product advisory council, which is a conduit to gather specific requirements and validate product plans. We will also be forming an executive board of advisers to gain insight and validation about our strategic direction.
How are customers reacting to the partnership? What specific results have you seen there?
Goodnight: Since last October, the SAS and Teradata CoE team has met with almost 100 joint customers and has participated in more than 50 customer engagements around the world and across all vertical markets. Customer feedback has been so positive that both companies are increasing global staffing to meet growing demand for consulting services and architecture assessment workshops.
Koehler: One leading insurance provider in North America was able to reduce its claims reporting process from 40 minutes to less than one minute through a set of SAS product enhancements, coupled with joint CoE recommended changes within the Teradata architecture to support SAS Business Intelligence. This performance increase is an example of what other customers have [experienced] and will continue to experience. Warner Home Video is another example where SAS and Teradata best practices for improved integration were leveraged. Taking advantage of its Teradata infrastructure, the company significantly enhanced its SAS run times. The model that previously took 36 hours to complete now runs in one hour and 15 minutes.