Sergey Nivens - Fotolia
Editor's note: SearchITChannel periodically looks into broad areas of technology to assess channel partner opportunities. In this feature, we examine big data and emerging roles for channel companies in such areas as big data consulting and big data platform deployment.
It's hard to imagine a mining company being digitally innovative, but that was the case with one such firm that discovered it could use data to anticipate when a massive piece of equipment might fail while in rough terrain at a remote location.
The company liked the idea of proactively performing maintenance on equipment before something went wrong, and wanted to take advantage of data being captured on thousands of sensors on hundreds of pieces of equipment at multiple mining sites, said Tim Brooks, big data principal consultant at World Wide Technology (WWT), which worked with mining company officials to set up a predictive analytics system to utilize the data the sensors were capturing. The sensors simultaneously gather data every few seconds and OSIsoft's PI System, a recording software product, logs the data, Brooks said.
The mining industry "is not exactly Amazon.com in terms of the cultural setting of most of the folks that have worked their way up from operating heavy equipment or running a mine to becoming senior executives of a company," Brooks observed. "So gathering that group to discuss how they use data and how they may derive value from it was a formidable challenge."
WWT examined and analyzed data that was gathered for 16 weeks in a Hadoop environment in its internal big data lab. Officials identified three instances that correlated with future engine failure with 90% probability on the heavy equipment, Brooks said.
This was a significant finding, he noted, because mines tend to be located in harsh, remote environments, and when a truck's transmission fails or an engine goes down, "it effectively shuts the mine down, thereby ruining productivity for a day, at a tune of one million dollars," he said. "If WWT can help its customer identify in advance which trucks are most likely to fail in 30 to 60 days, they can be taken out and proactively fixed."
This is a prime example of predictive analytics "applied in a very non-digital way in order to significantly improve performance and safety," Brooks said. A bonus finding was that WWT was also able to identify some truck operators who were not following company policy on speed and idling time, and those variables contribute to truck failure, he said. The mining company has since invested in its own big data infrastructure in a hybrid cloud model.
How channel firms can get their footing in big data consulting
Big data is, well, big, and that means opportunities exist for channel firms to establish a niche and help clients in a number of areas, like big data consulting, in the case of WWT, as well as deploying platforms, tools and offering managed services to support analytics projects.
Gartner estimates there are currently about 8.4 billion consumer and business connected devices generating data. This number is expected to reach 20.4 billion by 2020. But while it comes with opportunities, developing a big data practice requires a variety of soft skills and hard skills, industry observers said.
Tim Brooksbig data principal consultant, WWT
"I don't think it's difficult to establish a niche, you just have to identify what it is you're offering," Brooks said. Big data is "broadly applicable across verticals, so there's room for services, software, hardware and consultative providers, so you have to have some rigor and be ambitious and find your niche."
Channel firms can also help organizations understand what they need and what they can do with data science and machine intelligence as access to data continues to grow and more tools and technologies come into play, added Mark Jacobsohn, senior vice president and innovation service officer in the Strategic Innovation Group at Booz Allen Hamilton.
"Beyond the tools and technologies, there is a foundational component that will require cultural and organizational shifts," Jacobsohn said. "Considerations like ethics, security and governance cannot be an afterthought as we scale with data science and machine intelligence. Channel firms need to enable organizations to acquire all of these products and services in support of the booming data market."
Reseller Trace3 set its sights on partnering with Cloudera, Hortonworks and MapR as companies build data science pipelines and look to make prescriptive or predictive recommendations using licensed versions of the software, said Carey Moretti, vice president of consulting on data intelligence.
Big data can be utilized in most industries and it's up to channel firms to listen to what their clients' business needs are, she said. "Understand why your clients are implementing what they're implementing -- why are they buying 60 Cisco UCS boxes?" When clients make tech purchases, it could be tied to analytics or a data science or artificial intelligence (AI) initiative, she said. Or it could simply be to replace older technology. "Understand the true business reason IT is making a purchasing decision," she said.
From there, the partner can find an opportunity since the client may need help setting up a cluster or securing and installing software, Moretti said. "You have to have a passion for this space because it means having a lot of business conversations and understanding the ROI -- either for decreased cost or increased revenue and understanding their staffing needs."
Knowledge of a customer's business needs and processes and how big data can be utilized is critical, agreed Ashim Bose, director and product engineering leader of big data and analytics, at newly formed enterprise services integrator DXC Technology. "At the end of the day, knowledge of customers' use cases, workflows, processes, where big data would be leveraged, where it sits, how it needs to be consolidated and having intimacy with a customer is definitely a good thing for a partner to know and a good place to start," he said.
Enterprises struggle with how to rationalize and operate a big data platform, and often, they don't know how to bridge the gap between data outcome and data strategies to get to that outcome, added Dan Hushon, senior vice president and CTO at DXC. Partners can add value if they know how to use incoming data and real-time analytics, he said. People may know how to configure reports, he said, "but they may not always know how to translate business reports to get to a future outcome, and they're looking for big data skills on the business side as much as the tech side to get there."
While some smaller channel firms might think their clients aren't interested in big data projects, Moretti believes they may already be unknowingly talking about it or implementing them. "When you think about process automation and efficiencies, the change we're seeing is there's a ton of opportunities, and if something can be predicted or automated, that's where industry is going, and they will need a platform like Cloudera and Hortonworks for big data processing and analytics."
Big data in action
DXC is seeing three ways customers are using its big data platform: the first is in data discovery. Here, customers recognize they have a lot of data and perform some process to extract value from it and try to figure out how to do it quickly and in an ongoing manner, Bose said.
The next phase is analytic app development: Once there is an informed use case, the application is then tested to make sure it works correctly and further information may be added to improve the outcome, he said. Frequently, it is DXC that is asked to do the testing. The third phase is deployment and production. That is, operationalizing the app and embedding it in the customer's business and ensuring that the improvement is generating the necessary business outcome.
"We're seeing steady demand across all three phases," Bose said. "As customers are getting more mature with big data, the demand for the production environment is ramping up beyond discovery."
WWT has a big data proof-of-concept lab and offers hands-on workshops, which came about because customers were requesting deeper dives into infrastructures and more information on big data use cases, Brooks said. The half- and full-day workshops are tailored to a customer's individual needs and look at their current IT environment, their business objectives and how they should proceed with big data initiatives.
They are highly interactive and WWT provides subject matter experts and moderators. "In almost all cases that has resulted in great insights achieved and in many cases the workshops have turned into initiatives for roadmaps, investments in big data infrastructure and proof-of-concept testing," he said.
Skills needed and how to find them
Channel companies keen to offer big data consulting and other services will likely need to hire staff. It's no secret that people with data science, analytics and AI skills are hard to find and they can command high fees, observers said. Most said certifications, while valuable, are not as critical as the right cultural fit and desire to learn. One tactic is bringing in more junior staff who have "a ton of potential," Moretti said. "The key is finding folks that want to grow with you and give them a great place to learn."
Trace3 looks for people who are "humble, hungry and smart," she said, as well as those who want to be part of a team and are sharp and have the ability to work with clients.
That said, the company also looks for people who can code in different languages, which is important for utilizing all the tools that fit around the Hadoop ecosystem, she said. Other key skills are a good understanding of high-performance systems and the ability to troubleshoot platforms.
"Diversity is critical to building a successful cohort of data scientists," Jacobsohn said. "While it's important to have a mathematical, scientific or analytical background, it's equally as important to employ people with creative backgrounds."
Partners have to decide if they want to be "the general contractor building the house or the operator improving it over time," Hushon observed. "For me, the most interesting skills are people bringing domain analytics knowledge to the table, which tends to bring programming and analytic techniques and data visualization," he said. More advanced skills would be ML or AI knowledge: how to set up and tune and evolve neural networks and the ability to deal with correlation and causality for models for machine learning, Hushon added.
Both he and Moretti emphasized that traditional methods to obtain training no longer apply. "The old days of reading a book on a plane and being ready to go isn't going to work," Hushon said. There is free training available in almost every domain on Coursera, he noted.
"There are many technologies involved in these ecosystems, so you're not learning from books anymore," Moretti concurred.
Explore another perspective on creating a big data consulting practice.
See what other IT consultants are saying about big data
Learn how Singapore's central bank harnesses big data