Bill Ruh, vice-president of GE Software, believes machines are getting smarter. The company has expanded its vision of an industrial internet to an industrial cloud.
In 2013, Computer Weekly spoke to Ruh about a new software-enabled services business that the industrial giant was building.
The new business would provide predictive analytics based on continuous device monitoring on the so-called industrial internet.
At the time, Ruh discussed how jet engines and wind turbines could be monitored in real time to improve efficiency and reduce failure rates.
Two years on and he says GE’s business is now very much powered by software services.
Customers want outcomes, says Ruh, and from a GE perspective, these outcomes are realised through the idea of an industrial cloud platform. "We have been focused on building applications to generate these outcomes," he adds.
Ruh believes an industrial–based business should not ignore the threats and opportunities made possible when devices are connected to the internet. "Our biggest worry is if someone figures out how to manage a jet engine in a better way, which is a bit like when iTunes disrupted the music industry," he says. "All industrial companies need to create services using analytics data and sensors."
GE CEO Jeff Immelt believes the company needs to become a software business. In his 2013 letter to shareholders, Immelt wrote: "We will lead as the industrial and analytical worlds collide. We believe that every industrial company will become a software company; that, ultimately, deep domain knowledge is tougher to come by than writing code. In this sense, GE seeks to link the 'iron and the data'.
"Our assets have sensors that produce a customised stream of performance data. We can harness this data to optimise performance. We know that smart machines, guided by domain-based analytics in a distributed setting, will drive new levels of productivity. We call this the industrial internet."
GE is focused on building software. According to Ruh, this software added $1.2bn to the company's bottom line. But this does not mean GE will become an IT company. He says: "People don't think we'll build the next smartphone. We're not going to be a Microsoft or SAP. But any industrial business that does not become a software business will not be relevant."
Industrial internet built on IoT
Starting with a definition of the industrial internet, Ruh says: "The industrial internet is about using data from connected machines to make them more efficient. We are pulling data off our wind turbines and jet engines. On our jet engines, 5,000 data points are being generated continuously in flight. We are also putting cameras on our outdoor light fittings to enable a light fitting to provide more than just light, by having intelligence, which enables it to be used for applications such as traffic analysis."
GE has created an industrial application platform called Predix, on which it is now building applications. "We have created 40 applications," says Ruh. "One of these, PowerUp in our wind turbine business, uses sensors about weather and turbine data to develop best practices to enable operators to generate up to 5% more electricity without physically changing it, which generates 20% more profit for our customers."
Another example is an application to monitor oil pipelines. Ruh says: "We are building an intelligent pipeline system which is now being installed at Columbia Oil and Gas. We are combining sensor data and data from assets that Columbia Oil and Gas has in other locations to create risk models. This allows us to assess the risk of any problem occurring by predicting where they will see problems on the pipeline."
More articles on IT leadership
- Firms must incorporate agile into their strategic planning.
- Nobody has a monopoly on digital leadership.
- The digital CIO: Motivate the elephant.
- Interview: Building links between marketing and IT for the greater good.
- Four ways career healthcare CIOs are diversifying leadership portfolios.
- The digital CIO: Why embrace digital innovation?
- How CIOs drive innovation and create value across the enterprise.
- The digital CIO: The CIO is dead. Long live the CIO.
Ruh believes such technology will enable oil companies to build, repair and manage the two million miles of oil pipelines currently in operation, safely and more efficiently. "This is the idea of the smart pipeline," he says.
Growing the software business
In 2011, GE reported revenues of $2.1bn from its software, mainly licences in healthcare, intelligent platforms and the GE power grid line of business. Ruh says: "But we weren't really thinking about the big picture on creating greater value."
To achieve this, he says the company needed to develop an entirely new common platform and transform the businesses “to reimagine the service portfolio". This became the Predix platform, which Ruh created a team to build.
In 2013, this team was 100 strong; it has now grown to 1,000. Ruh says the team's focus is on analytics, cyber security and user experience.
Globally, GE has more than 10,000 software developers and architects. Part of Ruh's ambition is to connect these people to his software innovation centre and retrain them on the Predix platform and agile software development.
He admits one of the challenges he faced was to make the business comfortable with the concept of Predix as a software platform.
"But, in the end, we needed the platform to drive greater efficiency in our software development," he says. Also, customers expect integration, says Ruh. "We see some really specific patterns emerging of commonality in the industrial space, so we are inventing in these common application elements and putting them into the platform, so that you end up with 80% of the work already done."
Ruh took a strategic decision to build Predix on the cloud. The costs of IT infrastructure quickly add up – according to one of his recent tweets, 30,000 miles of oil pipeline generates 17 TB of data. "We realised that when we're talking about connecting machines, the cost to serve is high if you do it using traditional mechanisms. The cost of storage and building an application quickly and the cost of connecting a customer had to come down."
The only way GE could achieve this is through a cloud-based business model, mirroring the way the consumer tech giants use the cloud as a platform for their online services.
Challenges of becoming a software business
There are more standards than you can shake a stick at. We don't care about a standard unless it gets us quickly to a solution
Bill Ruh, vice-president, GE Software
The success of companies like Google and Microsoft has been fuelled, in part, by opening up their platforms to entice third-party developers. Similarly, open data is allowing innovative applications, such as for smart cities. But it has taken the IT industry a long time to see the benefits of opening up data and APIs.
This is one of the challenges Ruh faces as GE becomes more software-focused. "Many of our businesses say, 'don't you dare sell to our competitor'," he says. "But our software has to work with other products."
Industry standards may have helped the software industry to move away from locking in customers, but Ruh is not keen on standards. "There are more standards than you can shake a stick at," he says. "We don't care about a standard unless it gets us quickly to a solution."
Instead, he says: "We created a consortium focused on test beds. We want to test interoperability, so anyone can bring in their stuff and test it."
The consortium, called the Industrial Internet Consortium, now has 150 members.
Another challenge for Ruh is how to charge for GE’s new software capabilities. He says: "If you support 50 million devices where there are a few gas turbines and there are many valves, you still have to serve the same API."
Clearly, charging on a per-device basis does not take into account the differences between a simple valve and a fully operating gas turbine. "So you need to charge on a use basis," he says. "The main dynamic is the cost of a gigabyte of storage plus the cost to serve the API."
For Ruh, the value of the API needs to be based on the value of the outcomes it can achieve.