LAS VEGAS: While one could be forgiven for thinking that IBM Enterprise2014, which was themed as Infrastructure Matters, would mainly be about hardware, that was debatable, based on the initial product announcements. Fast forward to the end of the event, the bulk of the one-on-one briefings and the last formal announcements, and one could argue that the conference was primarily about analytics, Watson and its ecosystem.
As part of this week’s Watson (an AI system capable of answering questions posed in natural language, i.e. 2001’s HAL or Terminator’s Skynet) news, IBM announced the opening of its global Watson headquarters in New York City’s Silicon Alley, and five new Watson Client Experience Centers around the world. “Watson is bringing forward a new era of computing, enabling organizations around the globe to launch new businesses, redefine markets and transform industries,” said Mike Rhodin, SVP, IBM Watson Group, in a prepared statement.
“Watson is fueling a new market and ecosystem of clients, partners, developers, venture capitalists, universities and students. The next great innovations will come from people who are able to make connections that others don’t see and Watson is making possible.”
According to IBM, the potential ecosystem is immense. It said there are 19 million developers and a new generation of innovators entering the workforce, with entrepreneurs, startups, and enterprises looking for ways to mobilize their business, transform industries and improve society. This week the company unveiled a number offerings from the Watson Ecosystem, built on IBM Cloud with SoftLayer at its base, including patient care (human and animal), and vacation planning.
IBM announced in January it is investing $1 billion to establish the Watson Group (to develop and commercialize cloud-delivered cognitive advisors), including $100 million dedicated to venture investments to support start-ups who are building cognitive apps through the Watson Developer Cloud. Over 3,300 businesses and entrepreneurs have applied to the Watson Ecosystem, and in August IBM announced Watson Discovery Advisor, a system that can visually reveal patterns and pinpoint connections in data to accelerate the discovery process.
Analyst Joe Clabby called the recent roll out of Watson Analytics a “transformative event for the analytics industry” and also gives IBM a “huge competitive advantage” when it comes to dealing with Big Data databases. He said the company is years ahead of the competition.
“IBM has invested heavily in cognitive systems development, and in business analytics development and acquisitions for over a decade, and it will likely take other vendors years to catch-up. If IBM can convince businesses and public sector organizations such as schools to adopt this technology, Watson Analytics could and probably will become an unstoppable force in the analytics marketplace.”
Another analyst, Charles King, illustrated why hardware’s luster has dimmed in recent years. According to IBM at its recent System and Technology Group (STG) briefing, hardware shipments, which constituted 80% of total sales in the technology industry in the ’80s, had fallen to 37% by 2013. During that same period, the hardware portion of IBM’s revenues had nose-dived from over 80% to just 17% last year.
Even the hardware execs were pushing the analytics agenda at Enterprise2014. Mike Kuhn, VP and Business Line Executive, IBM Flash System, said the massive growth in data, and the need to analyze that data in as close to real time as possible is a major driver of the skyrocketing adoption of flash.”Flash is growing 10x faster than average storage… and is ready to displace high-performance spinning disk.”
He said it is starting to be positioned for the 20% of your data that needs to be on tier one storage; it’s already won the battle for tier 0 data. “We’re going to see disk quickly give way to flash in the performance tier”.
According to Cindy Grossman, VP, Business Analytics & Optimization, IBM Systems & Technology Group, the Power8, System Z and storage announcements centered around analytics. “Analytics is what clients need to grow their business. It’s all about business outcomes.” It’s still early days, she said, but analytics use is growing in IBM’s customer base, especially in areas like fraud detection and retail.
Customers have been telling Tom Rosamilia, Senior Vice President, IBM Systems & Technology Group, that they are swimming in data. “I don’t need more information, I need more insight,” he said during his keynote presentation.
The growing importance of Big Data and analytics will fuel a rebirth of the mainframe, said Ross Mauri, General Manager, System z, IBM Systems & Technology Group. The multi-decade trend of getting data off the mainframe into some distributed system/s appears to be ending, he said.
“I think that trend is going to reverse itself, keeping that data on the mainframe and doing the analysis real time”, he said. It’s no longer about the economics, it’s about getting them the data in a timely and effective matter.
He added that mobility and cloud are also helping to reverse the trend away from mainframes, but it mainly comes down to analytics. “That’s central one for me because of the data.” He only joined the Z team three months ago, but Mauri said IBM has been signing up 10-20 net new clients per quarter.
Bernie Meyerson, PhD, IBM Fellow and VP, Innovation, IBM, offered a number of interesting insights about analytics and innovation. He was more than happy to deflate the argument that there is a choice to be made between innovation and good-enough computing. “Innovation for innovation sake is stupid,” he said, but not innovating is even worse.
“The key to innovation is knowing the market you’re working in… (and) one of the great weapons we had over the years was we could do pure research without telling people where we were going.”
Disruptive innovation – i.e. the PC and mobile phone – has contributed to the declining life expectancy of Fortune 500 companies, now under 15 years. While Myerson wasn’t predicting that Intel would join that list, he did say that Moore’s Law – doubling the number of transistors on a chip every two years – will be history within the next 5-10 years. “The brute force approach won’t do it.”
Analytics based on alorithms and performed on general-purpose CPUs is a waste, he said. “The challenge is that the speed of light is way too slow for what we’re doing now.”
Last week’s announcement of the Power S824L servers, which can replace 24 x86-based servers, is a good start, said Myerson. In the future, by reducing the distance that has to be travelled between CPUs and memory from meters to micros will provide between 100,000 to 1 million orders of improvement, he said.
DISCLAIMER: In addition to being an IBM shareholder, IBM looked after airfare and hotel expenses for this event.