UPDATE: This story has been updated with a response from IBM.
NASHVILLE — OpenText CEO Mark Barrenechea this morning called out IBM’s cognitive learning engine for its closed ecosystem of deployment and development and urged enterprises to instead invest in his company’s future cognitive analytics technology.
Barrenechea made the case for OpenText’s artificial-intelligence based predictive analytics engine — to be called Magellan — over IBM Watson during his keynote speech at the Waterloo, Ontario-based enterprise information management company’s 2016 Enterprise World Conference here at the Gaylord Opryland Resort & Convention Center.
IBM 'Makes No Sense'
Here, in Music City, USA, the OpenText conference venue is just a short walk up Opryland Drive from the cathedral of country music, the Grand Ole Opry. Down the street, Barrenechea promised the crowd of 2,000 attendees and 300 partners to be “very direct” in his cognitive computing comparison to IBM, much like the nature of lyrics written by many famous country musicians who performed at the Opry.
“Our path on software is pure openness: open standards and open algorithms,” Barrenechea said. “IBM’s platform is closed. Our hardware is built open standard x86. You can chose your vendor, and you can choose your design. With IBM Watson, it’s closed. You use the Watson mainframe when you’re there.”
Barrenechea’s anti-Watson tune went on, as he criticized IBM Watson for owning all algorithms built off it while OpenText, he said, promises to allow developers to own their algorithms on Magellan.
“It makes no sense to me,” Barrenechea said of IBM owning algorithms. "IBM owns the algorithm. Not with us. You write your algorithm, you own it."
The OpenText CEO also said users can only run IBM Watson in the cloud vs. “anywhere” for OpenText Magellan: on premise, on the OpenText cloud or via a third-party host.
“We’re not going to place limitations on where you can run Magellan,” Barrenechea said. “It’s extensible by you where IBM is mainly extensible by its global services with Watson. Our ecosystem is open because of Spark Apache where by nature IBM’s ecosystem is closed. Most strikingly, Magellan will be highly affordable where Watson is astronomical.”
IBM: We've Got Open APIs
Reached by CMSWire after Barrenechea's comments, IBM Watson Chief Technology Officer Rob High said the cognitive APIs and services on the Watson platform are available to any developer who wants to build a Watson application.
"These APIs are available to the public on our Watson Developer Cloud," he said, "but we're expanding to other platforms as well."
He cited IBM's May announcement that it's bringing Watson services to the Twilio Marketplace, making them more readily consumable to Twilio's one million developers.
"We make these available in the cloud so they're more consumable for developers and to make it easier to get started without worrying about hardware or deployment," High told CMSWire. "Over the last 12 months, the number of developers using Watson APIs is up over 300 percent, and the number of enterprises we’ve engaged with has doubled."
Content Meets Analytics
Barrenechea produced a slide for the keynote audience that compared IBM and OpenText in the cognitive arena. It drew some oohs and aahs from the crowd and some immediate response on social media:
Holger Kisker, vice president and research director of BI, analytics and big data for Forrester Research, told CMSWire in an interview after the keynote that OpenText’s cognitive analytics future can be “great for the market and the community.”
It combines, he said, content and analytics exactly the way OpenText should be doing it.
“It fits in perfectly for them,” Kisker said. “They’ve made very clear they’ve put much more focus on analytics coming from their heritage with the content side. Everyone is wanting to make more insights and value out of content and combining analytics with content is exactly the right way to do it.”
OpenText still needs to catch up with other analytics providers but the acquisition of Actuate is helping, Kisker added.
But going after the giant of cognitive this early with Barrenechea’s salvo at IBM Watson?
“If you want to be successful in the market you need to have a story and a differentiation,” Kisker said. “As far as I understand, the main differentiation for OpenText will be its open community that it will bring to market. Their big bet is on community.”
Spark Apache Foundation
Barrenechea said OpenText Magellan will leverage Spark Apache data processing. “There are thousands of open algorithms,” he said. “We want to democratize cognitive. … This is built on open standards. This is really important.”
High said IBM is glad to see OpenText is leveraging Apache, "because we're fans, too."
"We support and offer extensions to Spark, and IBM contributed its open source machine learning technology, SystemML, to the Apache Incubator, and it's now known as Apache SystemML," High said. "Perhaps this will be one of the pieces they pull into their Magellan offering. We're also using Torch, Caffe and Theano — all of which are open source, including their algorithm libraries which are entirely open source."
Magellan, meanwhile, can attract tens of thousands, if not millions, of developers with Hadoop or SQL software running underneath the platform as it integrates with capabilities OpenText acquired through Actuate.
“It’s a major step forward for us,” Barrenechea said, “to bring Spark Apache inside of Actuate. We’ve thought long and hard about our next generation of analytics, and it is Magellan.”
All images by Staff Reporter Dom Nicastro.