Live from Yokogawa-- Mike Brooks from Chevron

Brooks, who is Staff Technologist for Chevron, described Chevron's global operations. The issues that drive Chevron's business-- --world market--crude supply--value in an "integrated energy company"--pursuit of efficiency with new ways to work, new technology deployment, realtime enterprise--aging workforce and technology complexity. At the highest level, it is quite simple-- refining=people and work processes. At the center is "sustainable business value, with business process integration, organizational capability and the proper enabling technologies intersecting. You have to have the work processes and the people processes first. At bottom, there is operations and at the top is the business system--and between them is Production Management. Manufacturing operations ranks number 1 for IT spend now, by lots. Here we have an opportunity to create a lot of value--and it is up to us to drag it out. Brooks described the work process for handling a problem. "It happens at the morning meeting, and the guy who shouts the loudest wins!" Shouldn't there be a way to manage and institutionalize those processes? We're talking about collaborative work processes which are auditable and event-driven. How do we do this in a more institutionalized way? He showed a sample work process-- the line up. You will not solve this issue with applications integration software. Remember ISO 9000? It was no guarantee of good quality, just great documentation. "It just guaranteed you'd make very repeatable low quality stuff," Brooks quipped. At Chevron we've created a Forum. We need collaborative work processes enabled by -but separated from--the IT underpinnings. We need secure data within primary systems but with easy data exchange and sharing and users need to find information easily without knowledge of where it is stored-- and we need to isolate and abstract implementation of vendor packages. This generally gets vendors upset, because it means that we can move much more easily from vendor to vendor, but that's what we want and what we'll get. The key ideas then: we are working on metadata. This industry is obsessed with tags. We want to move beyond tags-- you don't have to give the airlines your reference locator number anymore, you give them your name and where you are going. We don't need to see instrument tags, we want to see "what the level in the vessel is." We need plain language instead of tags. We need to have a federation of nameservices, metadata and directory services. We need to have a common exchange, and the system needs to become a broker of services. "I am a big proponent of SOA," Brooks said. And we need to have technology alignment. We need to work with Microsoft, for example, and they need to hear what they are doing that isn't what we need. We need a rich SOA that supports an IT foundation that can be flexible and enable work processes, he said. We need a different and more agile form of foundation IT architecture. At the bottom, it is a messaging system, and then you add event detection to it. This then becomes the control system. We need to have a data model (external model map, metadata, nameservices) that has persistence (intelligent cacheing, datastore, data warehouse). On top of that, we'll put the Business Process Services Execution Architecture. This will have Runtime Services (composition, business, application services) and Workflow Execution (business process model, roles and responsibilities) and Governance Services. This is an OPPORTUNITY (sic.) to remove the work process from the IT underpinnings. But it is only an opportunity if we grab it, and understand what it means. So what's it mean? We want to have a complete system (ecostructure) of reusable services that enable a nimble IT infrastructure that can adapt to meet changing business needs. From the plant floor to the business layer, Brooks indicated that this means we need a set of reusable interchangeable services. This is what we often call the "Lego" model of re-usable building block services. We need a clear SOA governance model that enforces policy globally, while stewarding integrity of the services provided. Brooks described an SOA Maturity Model. He noted that when he asks most vendors for reusable services, they start talking about wrappers around APIs. "I have news for you, wrappers around APIs is only just the beginning," he said. Chevron wants to leverage standards, because standards lead to agility. They produce lower TCO, cost sharing, agility, and interoperability. Brooks showed the BP Data Model Map with a standards overlay..."The good news," he said, "is that there's lots of standards. The bad news is that there's lots of standards." These standards are mostly reference standards. This means that OSIsoft and SAP might both tell you they conform to ISA95, but their implementations might be entirely different and incompatible... What we need is implementation standards. Brooks pushed the openO&M standard that is a consortium of Mimosa, ISA, and others, with open work process integration. "We want to map services once, and use them over and over in any refinery we run." He detailed the SOA openO&M proposal he's making to Chevron in a center circle-and-spoke diagram, and said, "This is the way to Truth. Everybody's truth is different, but this way we can give everybody the best shot at seeing the truth from their point of view." Chevron wants to lead the industry by driving COTS, using standards, and pushing hard for services maturity. "But it's got to be more than a bag o' services and workflows. It needs to be a repeatable framework on which we can build." Key ideas