Bentley Systems

Towards an open future: Keith Bentley interview

457 0

Following on from Bentley Systems’ Year in Infrastructure event, AEC Magazine interviewed CTO, Keith Bentley on a range of topics facing the AEC sector, software companies and customers, from virtual events, cloud and subscription to digital twins and the future of 2D drawings

We’re now into our second year of the main AEC software developers hosting virtual conferences instead of physical conferences. At the tail end of 2021, Bentley’s Year in Infrastructure event was broadcast from the company’s Exton-based headquarters.

The primary content covered the projects, customers and winners of the annual Year in Infrastructure awards.

Unusually there wasn’t a technical keynote from the company’s CTO, Keith Bentley. However, AEC Magazine did get an opportunity to catch up with him and discuss a wide range of issues that are currently on the minds of the editorial team.

AEC Magazine: With YII being run again as a virtual conference, how have you found the switch to online?

Keith Bentley: I miss sitting down in the same room with real users who say, “Hey, you’re not working on the right stuff ”, or “How is that new stuff you’re working on relevant for this problem?”. You just don’t get that feedback from people sitting in their homes, while staring at screens or listening to recordings. Virtual events make it easier to consume but it doesn’t have the same interaction value.

Once a year I would get ‘readjusted back’, to aim at the right target because I spent enough time hearing from customers, sitting at lunch with someone who says, “You guys are so foolish!”. That is the kind of feedback that is invaluable and for us, it hasn’t happened for two years.

We are winging it now. You could say that every one of our developers has a great opportunity to connect with any user because they’re just click away and you could schedule a meeting. But I just don’t think it happens, it doesn’t happen anywhere near as frequently as I hope it would. There are certain things that just don’t happen over ‘Teams’ calls. Virtual events have lots of value, but I don’t think they replace physical events. I very much look forward to more physical events.


While I don’t like the flying part of it, I do like being in some city where you’re disconnected from your normal routine. You have to a plan for it. So, I have to get everything done before that week. It becomes a stopping point. I like the fact that you have to disconnect to go to a conference. When I get home and I think about all the feedback I’ve gotten, you know it forms the next year strategy.

AEC Magazine: We are picking up lots of complaints about software companies in general from customers. Whether that be their software provider’s business model dictating when they upgrade, lack of development, changes to licences etc. The top 20%, the mature users are wondering what they are getting for these increases in fees, where’s the value? And I know from talking with publicly traded CAD companies there is pressure on to show similar profitability to shareholders that Autodesk is delivering.

Keith Bentley: Looking at desktop versus cloud I think it relates pretty well with people’s mindsets about their concept of ‘value’, what they get for their dollars (£).

If I sell you a cloud service, then there’s no way ever you could do anything unless you pay your bill, because I, the software vendor, turned you off? Whereas before, when you used your own licences, people would have the copy of the software that they owned on their desktop machine, and they could do whatever they wanted because it’s already paid for. So yeah, with desktop applications there was always a relationship with the vendor about upgrades and support, and you got new features, but customers felt like they at least had their destiny in their own hands.

I believe the answer is cloud when you want, but be able to pull your data out of the cloud. Take the program that understands and interprets that data, run it on the desktop, disconnect from the cloud

When I hear people talking about how great the cloud future is, I agree. There are now a lot of wonderful things we can do by connecting people together through a cloud service. But I put myself in the shoes of someone who says, ‘If my only choice is something that runs only on some computer that I don’t own – and unless I pay my bill (and I don’t know how much my bill is going to be next year versus this year) then aw, geez, don’t sign me up for that!’

When I hear other vendors talk about their ‘great new cloud based everything’, I say no thank you. I believe the answer is cloud when you want, but be able to pull your data out of the cloud. Take the program that understands and interprets that data, run it on the desktop, disconnect from the cloud. People get on airplanes, perhaps not as much anymore [because of Covid], there are disconnects. There are reasons why the computer on your desktop is a better computer than the one you rent in the cloud, and I say you should have that choice.

Microsoft has Office products for desktop and the cloud. Have you ever used PowerPoint in the cloud? I find the PowerPoint version that runs in the cloud unacceptable – it’s slow and there’s always these big, vague pauses whenever you work on something big. You’re sharing time with everyone else in the world with their Azure account.

If you just wait for that download to happen, and you have that the PowerPoint file on your local drive, you are not subjected to any internet outages; you have total control. I always use the desktop product, so I always tell everybody we have to make software that works really well when you are connected but has the possibility that you can disconnect. And one of the reasons I think it’s important for people to be able to disconnect is because they want to disconnect from us. If they want to say ‘Bentley, I don’t want to have a relationship with you anymore and that data that I created using your products, that’s mine, not yours’.

If this were to only work through a cloud service – and there are many examples of vendors who are creating products that are cloud only – what are those people [customers] going to do when they get upset with those vendors? Export the data in some neutral format? That’s unacceptable! Stop paying and you can’t use any of that software that you taught all your users how to create and consume this data.

I’m very, very adamant that our users won’t choose to use our products unless we give them the choice to stop paying us and nobody, no one, wants to hear that. We’re in business to make money. So, when I say the way that we will make the most money is by making it possible that people don’t have to pay us, nobody believes me, and nobody really wants to envision a future where users walk away from us. But I say they’re not going to walk away from us if we treat them fairly. They’re not going to walk away from us if we’re giving them the greatest value for their dollar. They’d only walk away from us if we stopped doing that. Giving customers a choice of being able to walk away from us means that at least they trust us to use our tools to create the most value that they’ll ever create.

AEC Magazine: There is still a lot of negative reaction to Digital Twins as a technology and, from what we can tell, a lot of confusion with what BIM is and what a Digital Twin is.

Keith Bentley: We could talk about Digital Twins as hype versus BIM. Back in the day, we knew what BIM did, everybody knew what BIM did: you create construction documents, you take your BIM file, and you use a BIM application, and you perform whatever operation you want. And then you’re done, and when the project is complete you archive the file.

When you talk about digital twins, you’re talking about something you’re going to run your infrastructure company on it. It’s not an application. A digital twin is not a tool for a specific type of user; it’s a theme that, I think, will form the basis for people’s IT departments. IT departments are going to be their digital twin departments, because there’s going to be no boundary between the data that is consumed in their digital twin application and the data that’s consumed in their business applications.

There is hype about digital twins everywhere and I limit the scope of my hype to be about infrastructure assets, large-scale things that have a fixed position on the earth. I’ll call that infrastructure, and not just public infrastructure. There are companies that run that and, in many cases, the infrastructure is owned by governments, so they’re even more concerned about future proofing this information system, where they can combine all their traffic management systems, the digital twin of a city or even, as you know, a region.

There are reasons why the computer on your desktop is a better computer than the one you rent in the cloud, and I say you should have that choice

People think today of the GIS system as being the centre of a lot of connectedness in government agencies, government business. In the future, the GIS is just going to be a part of their digital twin. For us to be relevant for a government, a large owner or even a contractor who wants to participate in that ecosystem, we have to be able to say to the customer, with a straight face, ‘If we don’t give you the most value, you can use your data and our software against us’ and believe that. This is the point you’re making about people getting mad at their lock-in. That’s what it is, that’s the reason Autodesk can charge more for a product. People probably say this about us too – they can sit there and say, ‘what’s your alternative? What are you going to do? What are your options? Is there a different Revit?

Suppose you’re a Revit shop. You’ve got all your data in RVT files and you’ve got all your training around Revit, so the cost to switch is way higher than the pain of your Autodesk bill and they know that. Autodesk has managed to make, much to my chagrin, Revit a standard, an accepted interchange format and people use Revit for lots of things today that there probably isn’t a good alternative for.

Fast forward 20 years from now, there’s not going to be that [one] tool – Revit or the equivalent Bentley tool. We can’t only fault Autodesk for this, because it’s all the vendors. I think there’s going to be a suite of tools and the tools will all look basically the same (or maybe not look the same but work the same) because they’ll be based on an open standard, something that nobody actually owns. Anybody can get the source code to it. The key topic around information software is data. Data drives everything. That’s why Revit is so immovable, because the RVT file.

AEC Magazine: There’s a lot of talk that, in moving to the cloud, there won’t be files anymore.

Keith Bentley: Think of what you have on your local drive and, as I said, I believe people covet the possibility of being able to extract their data from the cloud to have on their local computer to access it and use it for a period of time. At some point they’re going to synchronise their work, but what you have on your local disk is going to be a file.

It isn’t that there won’t be files in the future, but the boundary of what’s in a file and the format of the file, I think is going to all be governed by external forces. Vendors like us won’t necessarily have the ability to decide how or where to store that data, because there’s going to be one ‘open body’ that says what it should look like. What is that file format going to be today? I don’t think there is one. There’s nothing we could look at today and just use it.

AEC Magazine: What about USD with ‘BIM’ or additional AEC extensions?

Keith Bentley: USD has some virtue. It’s good but you can’t do engineering. You know a USD file is a mesh, so you could export to USD. It’s geometry, it’s a mesh, it’s visualisation and very good for that. Bentley uses USD, a lot of our users are exporting to USD. I think there needs to be a higher level, or different format. When you talk about an engineering model there’s a lot of data, such as analysis, as to why something is (as it is). You don’t put that in a in a USD file. So, there’s the connections between things, there’s relationships between things, then there’s the schematics that control why something is the way it is, and there’s lots and lots of properties that don’t show up in a visual view. Then of course there’s the 4D dimension, if you’re talking about a digital twin of something, you’re going to engineer it in phases, it’s going to change due to external influences over time etc.

Keith Bentley Bentley Systems
Keith Bentley

A static digital twin is an oxymoron. There’s no such thing. You must always consider time and all the formats that exist today, except for possibly one (I think you know which one – Bentley’s iTwin.js) haven’t been designed to anticipate that there’s change over time.

There’s a federation of data sources, as not everything always goes in the same location. A digital twin is not a BIM model. A digital twin is lots of other data streams, connected, such as IoT sensors etc.

You don’t put that ‘in a file’. A digital twin is an eclectic mix of lots of data sources. There are two things that are really, really important. One is you need to have a way to take a local copy of it so you can disconnect and have a session; the other is that we are going write lots of software that pulls data from everywhere and combines it.

People talk about the future of ‘all technology’ and how it’s going to be machine learning and the input to machine learning is data in some intelligible format. We are going to need lots of data and a platform and a programming basis to start from. Nobody wants to start from just Java script or JSON.

The evidence from the world today, from when we first started talking about this (about four years ago), leads me to believe it’s maybe going take a longer timeframe than anybody cares to admit, but it’s going to happen and is happening. The real-world examples we gave at YII were pretty good evidence that the theory is right and the uptake is real.

AEC Magazine: Because of the type of customers Bentley has – multi-nationals, governments, owner operators – there is concern that digital twins are not for firms such as architects etc?

Keith Bentley: Well, you are right. Our user base tends to be larger firms and have larger projects, so that tends to be our natural first focus, but if I thought we were working on something that was only applicable to large projects or large companies, I would be alarmed because I don’t want that to be our hallmark.

There are lots and lots of users for whom digital twins can be a huge benefit. It doesn’t have to be the size of a city or large firms. Think of water companies; are those big enterprises? I don’t think so.

Our go-to-market plan has evolved substantially from the beginning when we were just trying to make it. Nobody had ever heard of it before, and no one could understand the reasons behind it! So, we had to explain it in great detail. Nowadays you can get started with our iTwin products without spending a lot of intellectual horsepower understanding how it all works. I think it’s the greatest opportunity we’ve ever had to try to make our solutions more consumable in smaller bites.

There are many examples of vendors who are creating products that are cloud only. What are those people [customers] going to do when they get upset with those vendors? Export the data in some neutral format? That’s unacceptable

AEC Magazine: It’s not even straightforward to take a BIM model and repurpose it for a digital twin. BIM models are not really being developed in a way for them to be taken forward in a process, they are an end in themselves, to produce drawings.

Keith Bentley: I think this plays in to what is the future of files. Today we’re saying your digital twin originates from applications that were conceived before digital twins were a thing — Revit, AutoCAD and MicroStation etc. Basically, all of our products were developed in an era where files were the thing and the cloud didn’t exist. Basically, the applications had a ‘desktop mentality’ because that’s all there was. There was no concept that you could either store the data off the desktop, or that you could consume a service that wasn’t an application installed on your local machine.

I think the future is applications that are digital point applications. You were asking about what’s the long-term future of Revit? Well, I don’t know what it’s going to be, but I don’t think it’s going to be a desktop application only. I think there will be a desktop component of everything. As I said before, there will be always be a need, that when you disconnect you have a local copy of what you need, but you don’t have a local copy of everything. But you only get that local copy by connecting to some service somewhere that has the main branch and you synchronise your changes. It’s the way programmers have worked for years, so I’m 100% sure it’s got to be right because with our projects, it’s very distributed. We always have lots of branches; nothing is linear in software development.

I think engineering design and digital twin design will follow that model, quite nicely. But it’s definitely going to take a different generation of applications from the ones that exist today.

In my opinion, what I tell people, is the iTwin platform isn’t just a consumer of BIM data created externally, it is a generator of information. Long-term, I talked to users in our roadway design world and their projects are today all conceived around the workflows of drawings.

Contracts are issued based on the number of drawings; litigation is always around. How did the actual construction differ from the design? And the answer to that question is always ‘what does it say on the drawings?’. And I think there’s a future where there’s no drawings.

The drawings are artefacts. They come into existence when they’re needed. And then nobody needs paper, and nobody needs PDF. The PDF only saves trees. It doesn’t save time. It doesn’t save errors. Once you export into a system as a PDF, you now have the problem of what’s the vintage of this piece of virtual paper? And that’s where the value is. You can’t query why something was calculated that way, or get all the kinds of information flows that can happen when you talk about a database versus a report from a flat file.

If you don’t want to call it a digital twin, fine. But it’s something that takes into consideration the history of all the existence – what’s there today, why it’s there today, feeding back the traffic flow, do an AI study on it. There’s all kinds of workflows that should be different than the ones that exist today. But I think it probably takes a breaking point for people to decide to do this next project a different way. This is what we’re trying to show people and are already starting on that.

The day I started doing this, 35 years ago, the drawings were physical pieces of paper, and we always talked about how fast you could you issue a new drawing. 25 years from now, I don’t see that being the case, there will be no drawings on any infrastructure project.

Main image credit: PowerChina ZhongNan Engineering Corporation Limited


Bentley ‘Going Digital’ awards in infrastructure winners 2021

Winner – Rail and transit

Network Rail + Jacobs: Transpennine Route Upgrade (TRU)
Location: Manchester / Leeds / York, United Kingdom

The Transpennine Route Upgrade (TRU) is a GBP multibillion railway enhancement program designed to double capacity, reduce carbon emissions, and cut journey times on commuter routes between Manchester, Leeds, and York.

When completed, the 100-kilometre route upgrade will improve connectivity and provide economic benefits to the North of England. To bring together the large volume of data and disciplines involved, Network Rail tasked Jacobs with implementing a routewide digital twin.

Acknowledging that paper-based processes and Excel spreadsheets introduced unnecessary risk and inefficiency across the team, Jacobs used the Bentley iTwin platform with ProjectWise, ContextCapture, and other integrated applications.

Using the digital twin meant over 1,300 staff could track, contribute, and analyse design data and asset information in real time.

Improved access saved the team 20,000 hours in the first six months, worth an estimated £1 million. Overall, the digital twin will save approximately £15 million.

Bentley YII 2021
Image credit: Network Rail + Jacobs

Winner – Reality Modelling

HDR: Diablo Dam Digital Twin Modelling
Location: Whatcom County, Washington, USA

After the overtopping of the dam in Oroville, California, Seattle City Light initiated major safety reviews of its six dams, including Diablo Dam on the Skagit River.

To improve survey safety and efficiency, as well as minimise risks inspecting the 160-foot-high dam amid a global pandemic, HDR’s team was asked to provide aerial drone services to supplement physical inspections. They wanted to use the captured data to create a digital twin model of the structure.

They selected ContextCapture and the Bentley iTwin platform to create a digital twin, accurate within two centimeters, from over 82 million survey points.

The team could merge architecture, engineering, and construction data with artificial intelligence across the lifecycle of the structure, identifying current and future maintenance and repair needs to ensure safety.

The digital twin provides a single reference point for the owner to understand the structure and reduce project costs, while increasing surveyor safety and facilitating decision-making.

Bentley YII 2021
Image credit: HDR

Winner – Water, Wastewater, and Stormwater Networks

Companhia Águas de Joinville (CAJ): Contingency plan to ensure supply in the event of drought (Joinville-Santa Catarina)
Location: Santa Catarina, Brazil

The city of Joinville experienced its worst water crisis in 30 years. Responsible for supplying water and sanitation services to approximately 600,000 residents in the municipality, CAJ-Joinville Water Company initiated a project to develop a contingency plan to maintain water supply during worsening drought conditions. They evaluated three alternatives in the preliminary study. However, their initial solution yielded insufficient flow transfer, water shortages at the weakest points in the supply system, and a reduction in water transport efficiency. As a result, they needed to implement a more comprehensive study of the municipality’s entire network.

CAJ-Joinville Water Company used OpenFlows WaterGEMS to create a digital twin of the distribution system, modelling 285km of network. Using the hydraulic model to simulate a new contingency plan, they determined an optimal solution that guaranteed supply in the event of severe drought, while saving approximately BRL 4.5 million compared to their original proposal. The Bentley-based scenario improves flow efficiency, reduces pumping costs, and cuts energy consumption by 574 megawatts per year.

Bentley YII 2021
Image credit: Companhia Aguas de Joinville (CAJ)