GE Digital initiates a huge turnaround in its attitude toward software and Industrial Internet development. GE invested large sums to build a Silicon Valley presence for its software. Hired many engineers. Took its industrial software base up a notch or two with its Predix platform. Tried to build its own cloud infrastructure. The mantra—not invented here.
[Late Breaking News: I was wrong. There will be another Minds + Machines. San Francisco, October 30-31. That’s an expensive trip. Anyone want to fund me? 😉 ]
During the last Minds+Machines conference in San Francisco new CEO John Flannery, barely two months into the job, said that GE Digital needed to work more closely with partners. Soon thereafter came the axe.
That is the context for this major announcement (this one came from Microsoft, so within it may be a bit of its bias) of a partnership. Following report is based upon a media blog from Microsoft.
GE and Microsoft announced an expanded partnership, bringing together operational technology and information technology “to eliminate hurdles industrial companies face in advancing digital transformation projects.” GE Digital plans to standardize its Predix solutions on Microsoft Azure and will deeply integrate the Predix portfolio with Azure’s native cloud capabilities, including Azure IoT and Azure Data and Analytics. The parties will also co-sell and go-to-market together, offering end customers premier Industrial IoT (IIoT) solutions across verticals. In addition, GE will leverage Microsoft Azure across its business for additional IT workloads and productivity tools, including internal Predix-based deployments, to drive innovation across the company.
GE also plans to leverage Azure across the company for a wide range of IT workloads and productivity tools, accelerating digital innovation and driving efficiencies. This partnership also enables the different GE businesses to tap into Microsoft’s advanced enterprise capabilities, which will support the petabytes of data managed by the Predix platform, such as GE’s monitoring and diagnostics centers, internal manufacturing and services programs.
According to Microsoft, leveraging Azure enables GE to expand its cloud footprint globally, helping the companies’ mutual customers rapidly deploy IIoT applications.
The global IoT market is expected to be worth $1.1 trillion in revenue by 2025 as market value shifts from connectivity to platforms, applications and services, according to new data from GSMA Intelligence. Note: I find this a very interesting comment.
As part of this expanded partnership, the companies will go-to-market together and also explore deeper integration of Predix IIoT solutions with Power BI, PowerApps and other third-party solutions, as well as integration with Microsoft Azure Stack to enable hybrid deployments across public and private clouds.
Fourth in the series of posts as I digest all of the information I gathered at Hewlett Packard Enterprise (HPE) Discover 2018 in Las Vegas. This post focuses on use cases. Yes, people, there are people some in manufacturing and some not who are using HPE IoT and Edge computing for fun and profit.
First off, a panel assembled by Tom Bradicich, VP and GM IoT and Edge and Ph.D. entitled Intelligence at the Edge.
Nathalie Elad of Comcast- We are an aggregator of data from homes sending this data from local server to cloud. He is working with HPE on virtualization. No, it doesn’t collect individual family usage to sell to others (yes, it came up). But the company does need data to know how to channel bandwidth. The challenge-“we double bits every 18 months and need to flex up and down during the day.” Interesting stat—there used to be 3.3 devices per house, now may be 20 or even 30.
Tim Thai, Tesla- OT—IT is still a challenge. “The Edge is dynamic, wherever business sets up shop.” Regarding IoT, there are “Things” in manufacturing-control and sensors. They incorporate sensors in testing of technology in cars. Not to mention “there are a ton of sensors in a car.”
Philip Rostle, Alfa Romeo Sauber F1 racing, discussed F1 race car as the edge. There are lots of channels coming off the car during a race. They measure performance versus predicted. You think you have connection problems, he described connection in race as “variable”. Every car has a GPS. They track all cars in the race trying to predict status of the other cars. They run scenarios, analytics, quickly at the edge during a race to help determine strategy. Took “moonshot” server power to the edge so that they get maximum performance within the rules of F1.
In a special breakfast session, we talked with the CTO of the Ryder Cup and European PGA Tour. Think you know golf? Ever wonder about some of the stats that the TV announcers can quote during an event? Well, the tour requires a lot of data. And to get that data, they need connectivity. Golf is also an entertainment event. There are 50,000 spectators at the Ryder Cup. They all expect WiFi to access real-time information about the tournament.
First the data. Every shot has a dozen parameters to capture for every golfer. These are logged on the course. To connect, they use Aruba wireless networking devices. There are 30 switches and 700 access points. They collect 20K data points for scoring; 140K data points for other shot information. “Data drives insights that leads to performance for golfers.” They can track each golfer and also track spectator traffic patterns. An untold story, they lay 18km of fiber cable each tournament; ready to go for Wednesday morning and tear down beginning Sunday evening.
Mike Orr, director of digital transformation at Murphy Oil, uses Edgeline on oil platforms. He noted that his biggest hurdle was working with IT mostly due to its legacy software systems. He made this technology economics point—when oil went from $140 to $20, company laid off many workers. The only way he could get his work done was with technology.
I’ve already discussed the Texmark Chemicals “Refinery of the Future” use case, but I learned a few additional points at this conference.
Intel supplied streaming video analytics—used for physical security/monitoring, open gate for railway access allowed humans and critters into the site, monitored for exception to alert operators.
Deloitte is developing an IoT practice. It assembled an ecosystem including NI, Allied, ThingWorx, OSIsoft, SparkCognition AI for pumps. It also developed the operator dashboards for the project.
All together there were 12 partners in the ecosystem that completed the project that included predictive maintenance for two critical pumps and the video surveillance system.
HPE coordinated the entire project.
The insurance company was impetus to do something to upgrade the technology. Texmark kicked off the project by renting a party bus and taking 15 employees to the HPE IoT lab in Houston. They saw a demo of a pump with FlowServe monitoring and analytics. Employees discussed and picked the initial project targets—two critical pumps in the process plus the “video as a sensor” for the railway access. Getting early employee involvement was the key factor for successful implementation.
Hewlett Packard Enterprise (HPE) held its annual Discover conference in Las Vegas last week. It has made a sizable commitment to Internet of Things (IoT) and the Edge—areas central to my writing for the past few years. I am floating a number of ideas looking for feedback as I travel, and I’ll bounce some of those here later.
There is so much I learned last week beyond even what I wrote Monday about the new Edgeline computer. Perhaps the best place to start is with my latest discussion with Lin Nease, Chief Technologist IoT at HPE. This was a continuation of a discussion we began in Madrid last November and resumed at Industry of Things World in San Diego in February.
HPE’s power of compute at the Edge fascinates me. Even though my being in Las Vegas precluded being in Boston for LiveWorx, ThingWorx came up in many conversations at Discover. Nease said that ThingWorx (product and division of PTC) has been a good partner. Back to compute power at the edge Nease mentioned this power combined with TSN—Time Sensitive Networking, a new extension of Ethernet promulgated by IEEE.
Indeed, there is sufficient power in Edgeline that an enterprising developer could, for instance, accomplish the software defined DCS that seems to be the dream of some of the engineers at ExxonMobil and the Open Process Automation folks. Anyone out there have time and money?
Speaking of Edge, evidently the enterprise IT bloggers I hung out with during the event try to avoid the term. CEO Antonio Neri had said, “Edge is everything outside the data center.” In the blogger round table that I posted Monday, blogger Alastair Cooke noted, “Gary, we consider everything you do as edge.” Back to Neri who stated 94% of data is wasted; 75% of data comes from the edge.
Following are some points I gleaned from a session called “Harness the Power of Digital Platforms”:
HPE is a huge fan of open source & open platforms
Digital natives build platforms-e.g. Uber, Google, Amazon, etc.
An internal team built an open API platform to solve a problem in supply chain
Biggest problem was selling the system internally so that people would actually use the system (never seen that before—said no one anywhere)
Traditional—>Digital; everything is a frictionless stream of data
Platform always on, always looking for exceptions — sense/respond
HPE has an OEM Solutions group. Following are some points from a session discussing them:
OEM Solutions can be Embedded, Integrated, Private Label
Everything as a Service — Green Lake is the service offering that OEMs can resell the service
Shift to software defined
From storage to flash
Example—Konica Minolta embedded an Edgeline computing device in a printer called workplace hub that makes it easier to set up and install a new remote office
HPE has momentum in IoT and edge devices—and an organization supporting manufacturing.
I attended the Hewlett Packard Enterprise (HPE) Discover conference as the IoT blogger. It is a different program from that of the press. More blog posts coming when I get a chance to catch up. Actually this week I’m in Florida at yet another conference. And I have things from two weeks ago. Jeffrey Powers with Geekazine live streamed and recorded the blogger / influencer sessions. This video shows the blogger roundtable that I participated in. It begins at 3:50 into the video.
Antonio Neri, CEO and President of Hewlett Packard Enterprise (HPE), used the phrase “Data is the new currency, memory the new gold” in his keynote to the company’s annual US customer conference Discover in Las Vegas in June. Just one of the many places I’ve been lately.
If you haven’t planned for data in your machine and process control designs, you had best begin.The race for improved operations performance is on now.
We talk often of “edge” in the world of Internet of Things or Industrial Internet of Things. The edge has many definitions, but it can be defined as any place outside a data center. PLCs, for example, not only perform logic control, but they also aggregate data from perhaps thousands of sensors. SCADA devices and industrial computers also collect and channel data from a few to many sensors and data sources.
Business operations managers are hungry for this data to feed their information systems that in turn fuel their business decisions. Data in context is information. Information correctly presented to decision makers leads to better, faster decisions—and a competitive edge.
This search for competitive edge has moved me from an emphasis on control and automation (something we still need to do well) to Industrial Internet of Things. The IIot is taken by many as a similar strategy to Industrie 4.0 or Smart Manufacturing or whatever different countries call their strategies. This means I’m looking at a new generation of edge computing, enhance networking standards, human-centered design for mobile visualization of data, and even Augmented Reality (AR) and Artificial Intelligence (AI). These are not far-out technologies any longer. They are here and applications are growing.
Neri talked about the future as edge-centric, cloud-enabled, data-driven. He said the edge is where the action is, where the data is created. HPE is going to invest $4 billion in the intelligent edge over the next 4 years.
The company announced a new edge computing device with enterprise grade computing power (far beyond a PC) plus up to 48TB (yes that’s Tera not Giga) of memory. Oh, and it also comes in an environmentally hardened package. The CTO of Murphy Oil talked of using these on off-shore oil rigs.
Texmark Chemicals is a Houston, Texas based petrochemical refiner. I had several opportunities to talk with them about their IoT projects. They orchestrated an ecosystem of 12 suppliers initially to instrument critical pumps in their process in order to achieve predictive maintenance. This potentially saves the company millions of dollars by avoiding catastrophic failure. (Note: I previously wrote about the Texmark use case here–and expect more to come.)
Back to the announcement from HPE about the new edge product—a family of edge-to-cloud solutions enabled by HPE Edgeline Converged Edge Systems to help organizations simplify their hybrid IT environment. By running the same enterprise applications at the edge, in data centers and in the cloud, the solutions allow organizations to more efficiently capitalize on the vast amounts of data created in remote and distributed locations like factories, oil rigs or energy grids.
(Dr. Tom Bradicich wrote a blog post you can find here.)
HPE’s new edge-to-cloud solutions operate unmodified enterprise software from partners Citrix, GE Digital, Microsoft, PTC, SAP and SparkCognition, both on HPE Edgeline Converged Edge Systems – rugged, compact systems delivering immediate insight from data at the edge – and on data center and cloud platforms. This capability enables customers to harness the value of the data generated at the edge to increase operational efficiency, create new customer experiences and introduce new revenue streams. At the same time, edge-to-cloud solutions enabled by HPE Edgeline simplify the management of the hybrid IT environment, as the same application and management software can be used from edge to cloud.
“The edge is increasingly becoming a centerpiece of the digital enterprise where things and people generate and act on massive amounts of data,” said Dr. Tom Bradicich, Vice President and General Manager, IoT and Converged Edge Systems, HPE. “Our edge-to-cloud solutions help bring enterprise-class IT capabilities from the data center to the edge. This reduces software and IT administration costs, while accelerating insight and control across the organization and supply chain.”
HPE also announced the HPE Edgeline Extended Storage Adapter option kit, adding up to 48 terabytes of software-defined storage to HPE Edgeline Converged Edge Systems. This enhancement enables storage-intensive use cases like artificial intelligence (AI), video analytics or databases at the edge, while leveraging industry-standard storage management tools such as Microsoft Storage Spaces, HPE StoreVirtual VSA, and VMware vSAN.
The Industrial Internet Consortium (IIC) generates much useful information promoting awareness and technical tips about, well, the Industrial Internet of Things. Last week I had the opportunity to speak to the authors of a new white paper, ”Introduction to Edge Computing in IIoT”, Todd Edmunds, Senior Solution Architect, IoT, Cisco, and Lalit Canaran, VP, SAP.
The paper provides practical guidance on edge computing, architectures and the building blocks necessary for edge computing implementations. The IIC is also planning to release an Edge Computing Technical Report that will contain in-depth technical information in the coming months.
This paper is not a C-level generic paper evangelizing the concept, but rather practical advice designed to open the discussion followed by technical details targeted to those to whom the C-level executives might tell, “I have been reading about the IIoT. This looks like something we should be jumping into.”
We discussed how the edge should be defined by the business objective rather than the technology used. Using computing at the edge improves performance of the system when bandwidth could be the constraining factor for using the cloud.
As the edge gets more powerful, they told me, the role of the cloud will shift to one of orchestration of remote sites plus storage.
“Many companies are wanting to realize the business benefits that edge computing is purported to provide but are unsure where to begin or how to realize those advantages. The IIC has been at the vanguard of the industrial internet since its inception, and edge computing has been an integral part of driving the transformational outcomes that go along with it,” said Edmunds. “With the publication of this white paper, we provide practical guidance on where the ‘edge’ is and the key drivers for implementing edge computing. We also provide detail on edge computing architectures and real-world use cases.”
“Almost every use case and every connected device on the industrial internet requires some sort of compute capability at its source at the edge,” said Dr. Mitch Tseng, Distinguished Consultant, Huawei Technologies, and co-author of the white paper. “Oil rigs in remote locations have sensors gathering data but they need to be mindful of the challenges of data transmission because of bandwidth issues or the cost of transmission. The white paper is a first step in the development of an industrial grade ‘cookbook’ for edge computing.”
“Organizations adopting an IIoT strategy need to understand what data is available, how to use it to drive industrial processes and how to orchestrate, manage and secure data/compute,” said Canaran. “This paper and subsequent technical report will enable enterprises to unlock the full potential of the edge-cloud continuum and drive the business outcomes enabled by next-generation IoT devices, machine learning and AI.”
The full IIC Introduction to Edge Computing in IIoT white paper and a list of IIC members who contributed can be found on the IIC website.