At Autodesk University 2021, the main hype was around Autodesk Forge, the cloud-based development platform for developers and customers. Launched in 2015, it’s been a long-time in the making, and it would seem the company is betting the farm on it. Martyn Day reports
At Autodesk University 2021, the company’s CEO Andrew Anagnost dispensed with pushing Autodesk’s core design and collaboration products. Instead, he, along with the executive team, spent time eulogising Autodesk’s cloud services, connectivity tools and the development platform that enables this, which is called Forge.
Autodesk has been building the Forge SaaS cloud platform for years, but until now it’s mainly been marketed to developers. Because the Forge development suite is so different to anything else on the design market, it needs careful explanation. Autodesk now needs to introduce its customers to new ways of thinking about design tools, data and how they should expect the Autodesk development community to plug into their digital ecosystems.
Covid certainly forced AEC firms to sink or swim and pivot their entire IT systems to serve more distributed users, scramble to buy workstation-class laptops and accelerate cloud-based collaboration tools. Sure, in this process, many firms found new bottlenecks and had to rapidly deploy a whole new way of working but, amazingly, the industry managed.
Autodesk feels this baptism of distributed working fire has forever changed the way firms work and now appreciate the flexibility and resilience of design firms having a cloud back-bone.
In his keynote, Anagnost wanted to get customers to look beyond this conversion to cloud adoption; stage one of which is the distribution of files, “We’re going back to work. But we’re not going back to the way we used to work, we’re certainly not going to stop storing our data in the cloud or collaborating in the cloud. If anything, we’ll be doing more of that. But when exporting, uploading and sharing giant files doesn’t produce giant returns, maybe it’s time we face a clear reality. It’s not our files that are valuable, but the data that’s locked inside them.”
Autodesk’s advice for leveraging the value from latent customer data is to utilise its Forge platform. Up until this point, Autodesk Forge was mainly talked about in the context of developers and APIs (Application Programming Interface) for cloud applications.
Turning the promotion of Autodesk Forge to customers seems like a big jump for developer tools as, I would suspect, very few users would currently have the ‘programming chops’ in-house to utilise it. But clearly, much of the messaging from AU was explaining to customers how Autodesk will be delivering a phase change in accessing authoring tools, creating data, collaborating around data, and applying technologies like machine learning.
Autodesk is certainly the most clouddevout out of all the CAD firms. As others mainly focus on cloud servers and collaboration / connectivity, Autodesk is looking to completely move the centre of design away from distributed, desktop, file-based networks to a cloud-centric common data environment. Anagnost refers to these as information models.
A Forge-centric future isn’t just a bunch of APIs and tools, it means an end to filebased working as we know it. Anagnost succinctly expressed how he feels about the potential of Autodesk Forge, “No matter what industry you’re in, no matter what products you use, or where in the world you are, making data more accessible, more extensible and more open will help unleash your talent, connect your processes, automate your workflows, and unlock valuable insights.
So what is Autodesk Forge?
In the old days of programming, applications such as AutoCAD were one big monolithic lump of code, comprised of millions of lines. AutoCAD was written like this up to R12. In creating the R11 to R12 update, the Chief Technology Officer (CTO) at the time, John Lynch, once explained to me how editing and updating that new release, using this procedural programming methodology, was getting harder and harder and taking longer to do. It was becoming a ‘rats nest’ of code.
When it came time to create AutoCAD R13, and facing the pain of refactoring the old code and now a port to Windows, the decision was made to redevelop AutoCAD and modernise the code to components, using object-oriented C++. This made adding and removing new capabilities and enhancements much easier.
Autodesk stumbled by releasing R13 too soon. But, once stabilised, subsequent AutoCAD releases benefitted from having better underlying structured and managed code. AutoCAD updates subsequently came out more regularly. For software firms, making these kind of architecture changes for products with millions of users is like changing a tyre at 100mph. Code maturity and inevitable underlying platform changes, lead to big decisions for software firms.
Fast forward 20 years and Autodesk now has multiple applications in multiple disciplines, multiple platforms, multiple code bases, multiple file types – all requiring dedicated development teams, each tending to their unique products.
Projecting forward, it was clear that the cloud would play an important part in delivering software. Mobile computing was becoming increasingly widespread and, while today’s software can utilise the Internet as an exchange server, monolithic file-based paradigms are a hindrance to distributed computing.
If you took all Autodesk applications, in-house or acquired, there are common capabilities, such as read and write DWG, viewing 2D and 3D models, manage changes to designs, render, generate CNC, point cloud support, file translators, PDF output – the list goes on. If desktop software needs to be refactored for this new cloud platform, then why not break down functionality to universal services which can be called on, in almost a plug and play manner?
New products could be assembled from standard common components, from across Autodesk’s entire portfolio of capabilities. Autodesk’s developer community could move to creating applications which utilise actual Autodesk services within their next generation products, saving them time and ensuring compatibility.
Autodesk Forge is a set of these core discreet subsystems with APIs which can be wired together in a myriad of configurations to provide core functionality. No more reinventing the whole wheel as all products will mainly be made from a kit of standard parts. Enhance a Forge component with new capabilities, and all applications built with that get that update.
When Autodesk now makes any acquisition, the proviso is that the application will be reworked to utilise Forge components and unique new functionality will be used to create new services. An analogy would be the Borg in Star Trek, the ‘Forgeification’ of CAD applications assimilates and standardises all the design functions Autodesk creates or ingests on its forty year journey. We can see this with Plangrid, Assemble Systems and expect to see Spacemaker’s technology also added to the Autodesk Forge stack for more widespread use.
As Forge has been in development for quite a while, there are numerous AEC applications which utilise a number of common Forge components. Autodesk’s new digital twin platform, Tandem, is a great example. Here the team rapidly developed a new application for an emerging vertical market in just 18 months.
Using pre-developed capabilities such as the Forge Viewer and API access to Autodesk Construction Cloud and BIM 360, it meant Autodesk could get alpha versions of Tandem in the hands of customers exceptionally quickly, gather their feedback and start building deeper layers of capability. In the same digital twin market, Space Group in Newcastle upon Tyne created an exceptional digital twin platform called Twinview, mixing its own cloud developed software with the Autodesk Forge Viewer and the Autodesk BIM 360 API.
AEC Magazine also features GeoBIM from Esri, the first cloud to cloud collaboration between software giants, which utilises Autodesk Forge components. The application enables users to combine information feeds from Autodesk Construction Cloud or BIM 360, with GIS data served up by Esri’s ArcGIS servers.
Users don’t have to upload any files, just simply select data from hosted projects to see the resultant co-ordinated combination of BIM and GIS in Autodesk’s Forge Viewer.
This approach to Forge and application development also means new business models for Autodesk, its ecosystem of developers and potential competition.
In the CAD software market, nearly all CAD developers protected who could and could not develop on their desktop systems. Many times, popular AutoCAD applications were bought by competitors, and they were quickly stripped of their accredited developer status. With Forge, combined with the business outlook of Anagnost, cloud-based design data changes the rules of engagement, as he explains, “We don’t really care who builds [using Forge], people are going to build things that compete directly with some of our products and we’re OK with that, because one of the things that is super critical here is that this move from files to APIs has to catalyse a greater movement inside of the industries that Autodesk alone cannot solve.
“It’s going to require a lot of people coming together and moving the industry to different ways of thinking and working so look for that ecosystem to grow and grow and grow over time. More APIs and people will be able to do more powerful things.”
For a whole industry that literally has a business model based on trapping customers in their proprietary silos, this will be met with a healthy dose of scepticism, and it will be a case of action speaks louder than words.
Openness has become a common theme for Autodesk and the definition of openness could be debated. But Anagnost was emphatic, “We’ve reached out to embrace all the important standards. IFC, obviously is something we’re leaning pretty heavily into. We believe this is important. We don’t believe open standards can do everything, they never do, but they’re critical to catalysing the transformation that I’ve been talking about in the industry.
Autodesk has an ambitious cloud strategy. It involves moving its customers’ data from local to hosted, going from files to databases, desktop apps to client, file exchange to API, proprietary to open, all while maintaining existing tools, legacy formats and working methodologies
“Without the support of open standards, we can’t get the whole industry wrapping around what the future looks like, and how these workflows should work in the future. It’s critical that we’re open, that we support open standards, that people get more APIs of the data that we create, so that they can do things with them. Otherwise, we’re not really going to accelerate this transformation. It’s still early in the journey. We’re not done yet, but we’re making lots of progress.”
Anagnost also explained that he knew open APIs would mean competitors would ‘consume data from his world into their world’ and he said he was OK with that. If they had reciprocal open APIs, Autodesk will talk to them. “We will always be living in a heterogeneous world. So, the more we embrace the APIs, and the more we embrace the connectivity, the better off we’ll all be.”
Anagnost doubled down, “We’re not going to stop anybody from using the environment, the way they are, we’re just determined to be better than them.”
In its first and most common incarnation, IFC is a file-based format, and historically a very low common denominator. It has improved with each revision, but still can be the source of great frustration.
Autodesk’s implementation of IFC in Revit developed a negative reputation, but last year Autodesk licensed the Open Design Alliance’s (ODA’s) IFC libraries which, once implemented, should hopefully raise its game.
The bigger issue is the potential trap of being owned by a software vendor. With all your data in their cloud, is it easy to leave? In the future we will probably be less worried about file exchange and more concerned about depth of integrations between AEC ‘cloud islands’ of data and capability.
While there is a big issue about openness between competing products, Autodesk with its historic ‘code of many colours’ had data interoperability problems between its own applications. With industries like manufacturing and AEC converging, not being able to connect designs between important disciplines is not going to be an option.
Here again, Autodesk is looking to use Forge and standardisation to remove its in-house silos of design data. Amy Bunszel, executive vice president of Autodesk’s AEC design solutions stated that the company was using Forge to build fluent workflows between Autodesk Revit and Autodesk Inventor, the mechanical CAD tool. “With Forge we are unlocking Revit data for use in nonAutodesk applications. Just like data flows between Revit and Inventor, soon Revit data will pass directly to Microsoft Power Automate, making BIM data available for a wide variety of uses, giving granular visibility into projects and making it easier to supply up-to-date information to partners without having to pass files back and forth.
If an object parameter changes in a model, a supplier won’t have to dig through a huge file to find the change. Instead, designers will be able to create an automation to instantly give their supplier the exact information they need. That’s just one example. But with Revit data unlocked, you will not be limited to the constraints of a proprietary file format.”
Customers & Forge
With the target audience for the Forge message at Autodesk University being customers, it was intriguing to hear Autodesk pitching Forge as a consumerlevel tool. Looking back, AutoCAD eventually supported a plethora of APIs (Arx, Lisp etc.), but few customers went beyond creating simple scripting macros. Forge would surely only appeal to a minority?
Anagnost responded, “In terms of how many customers are going to natively build off Forge, it’s going to be 20% of customers. There’s a lot of them actively building right now. But that’s not the point to Forge necessarily. We’re just going to underpin these vertical environments that we have facing the various industries.
“As a result, it’s going to allow us to automate and create transparency between various disciplines and designs in a way that you just can’t see today. And it’s going to allow a new ecosystem of third-party developers, because the old ecosystem has faded away in many respects.
“Thirty years ago, AutoCAD spread all over the world because of its developer ecosystem. Since then, that ecosystem has kind of plateaued. It’s not growing and changing like it was before. But this new ecosystem that we’re going to create through Forge APIs, it’s going to be, I believe, incredibly vibrant, and it’s going to create an opportunity for our partners, for new entrants, for customers to do things they haven’t done before.“
Anagnost envisages the majority of users will benefit from Forge through his company’s in-house developed services and add subscriptions to the services of their third-party developers. For example, in the case of digital twins, customers would subscribe to BIM 360 or Construction Cloud, and they would have a choice of using Autodesk’s own Tandem or Space Group’s Twinview. For large firms who want to create their own workflows, they can have their base subscriptions and then craft their own applications and plug in their own data tools using Forge APIs.
It’s going to be interesting to see how Autodesk decides to separate out functionality, bundling in these future Forge replacements for today’s applications. There is a danger that some customers will feel concerned that Autodesk is expecting them to ‘finish off’ the applications they bought, in a kind of Mr Potato Head way. These thoughts lead on to what will be the changes in the business model and cost?
Autodesk is renowned for evolving its business model. We went from perpetual licences to suites, to subscription, collections, named user licences, loss of network licences, pay per use. These have all been pushed through the user base, leaving many in license hell, running multiple products on multiple licence types. All this time Autodesk has known that its ultimate objective has been to host customers’ data and deliver apps and services through a cloud-based system.
Autodesk’s transition to cloud is going to take many years and customers do not like to feel as if they’re being herded in a direction where they feel they have less control or choice
Anagnost was asked how customers would access Forge capabilities, “Sometimes pay (per transaction), sometimes simply available to [under] the umbrella of a particular subscription. The idea is the customer is going to get to choose how they access some of these things, but we’re definitely going to bring capabilities to whatever you need and those capabilities will range from rendering, scheduling and all things associated with that.
“What we won’t do is sub divide them so much that customers are pulling up different environments and apps for different things. As we move into the future, certain capabilities will be presented to them as they need to work through things. Some will be available simply under the umbrella of their subscription. Some people will choose only to use our products and the pay per use mode because that’s what they want if they only use it once in a while.
“Our goal is to offer maximum flexibility to our customers with regards to how they access and use these capabilities, by making it easier for them to find what they need to do their jobs. Not have to sift through dozens of products, but log into an environment, start getting their work done and often, in many, many cases, be presented with what they might need in order to solve a particular problem.”
How pricing might change in the future for customers is a little fuzzy, but Forge enables much more transactional charging. If your data is held in Autodesk’s cloud, and you want to run generative design, rendering or some kind of structural analysis, Autodesk will now be able to charge per individual calculation or result. Of course this brings to light Autodesk’s token system, which will be used to charge customers for services beyond their subscription level.
Third party developers have already been using Forge to create applications. As the platform is transactional, fees can be due for using certain APIs and free for others. At the moment, accessing BIM 360 data is free, while some of the other APIs, such as the model derivatives API, will charge the developer credits each time a model is sent to be processed, and will vary depending on the complexity of the model.
Other free to use APIs in Forge include the highly popular viewer. However, as models need to be converted through the model derivative API to get a native SVG/SVG2 format, someone will pay for that transaction, either when the model is viewed or uploaded to BIM 360. Micro transactions are going to be an important part of Autodesk’s business model in the cloud.
Some developers who have looked at creating Forge applications which require a high level of model transactions and processing through certain Forge services have told me that it can get prohibitively expensive with scale very quickly.
For IT design managers, centrally hosted data and applications hopefully nix the current hassle of managing licences in an enterprise. However, the nature of transactional micro payments could prove to be an issue in tracking and monitoring usage. I hope Autodesk has a managerial system to enable customers to stay in control and cap spending.
There is no doubt as to the importance to which Autodesk is placing on its move to the cloud. It involves moving its customers’ data from local to hosted, going from files to databases, desktop apps to client, file exchange to API, proprietary to open, all while maintaining existing tools, legacy formats and working methodologies.
It’s going to take many years for this transition to happen and customers do not like to feel as if they’re being herded in a direction where they feel they have less control or choice. There was some reticence about moving to cloud-based storage prior to Covid. That quickly evaporated when offices shut. The next phase is to educate and make customers aware of what kinds of things are possible.
The current challenge for Autodesk will be trying to maintain its ageing desktop products, which will be absorbed into the Forge platform. Autodesk still needs to persuade customers they’re getting value for money with the updates they get.
As Autodesk has clearly stated it intends to eventually replace the current generation of desktop products with more dynamic, thin client style applications, I suspect that some of this user frustration comes from hearing the Forge utopia, but having no explanation or timeline as to how Autodesk plans to migrate from the current applications to Forge. Anagnost’s ‘don’t fixate on your authoring tools brand’ message is very significant.
The AU focus on Forge felt like the start point of that conversation and reading between the lines we are a couple of years out before there will be much to show in the AEC space.
With no next generation of Revit confirmed and a concentration on construction, it’s going to be very interesting to see how Autodesk defines its future AEC-focused information model and workflows.
If the problem is tackled with a construction centred information model, as opposed to design driven, this could be an issue for architects. The industry is crying out for expanded capabilities beyond the current limitations of today’s BIM tools.
The initial concept of the transition to Forge was a huge call. Autodesk’s competition is taking very different routes. I am sure many customers will transition without really thinking about it. Some may be concerned about Autodesk not just owning their design tools but also their process and all their supply chain / ecosystem. Wherever data authored by its customers goes, the micro transactions will be generating revenue for Autodesk.
While a Forge-centric environment is still some years off, AU this year gave users a picture, painted with words, of the direction of its development and what the company envisages for the future of design tools. It also produced some definitive statements on Autodesk’s intent to be open, which are welcome and, until proven otherwise, have to be taken seriously.
For twenty years the BIM market has been suffering a lack of interoperability. Now Autodesk, Bentley, Trimble and Nemetschek have all stated they embrace open formats. It will be interesting to see how portable customers data is and should they wish to leave a Forge platform, what technical challenges they would face.
The origins of Autodesk Forge
Computer science usually advances through evolutionary phases. Examples of the evolutionary approach would be updated Operating Systems (Windows 3.1 to Windows NT), increases in the number of transistors in CPUs, or hardware refinements such as Solid State Drive (SSDs) over Hard Disk Drives (HDDs).
Small to medium technology changes which occur year to year might not seem significant but, over five years, looking back we can see progress. Looking inside any computer, you are just seeing the latest instalment in open warfare between thousands of hardware and software developers to go faster, do more, to be the best.
It’s always amazed me how we grow and adapt with this computer arms race, changing the way we work and the structures of how our data is organised. Who would have thought those early 2,800 baud modems would eventually revolutionise phones, shopping and allow us all to work at home during a pandemic, while Boris Johnson partied, and others tracked down online supplies of horse worming pills?
However, less frequently, computing will see generational change, in which something happens which turns existing paradigms on its head.
In CAD, the move from expensive UNIX systems to DOS / Windows PCs provided an opportunity for a startup called Solidworks to create a solid modelling application which undermined the highly expensive and dominant UNIX-based CAD software developers. Like the dinosaurs, some became extinct.
The Autodesk Forge platform has been out in the open for seven years and has its origins in the development of Autodesk Fusion 360, which, amazingly, was launched back in 2013. Fusion was Autodesk’s third major attempt to create a product design system for manufacturing (its first being Mechanical Desktop, based on AutoCAD, and then Inventor).
In manufacturing, Autodesk was always, volume wise, fourth behind the main pack of Dassault Systèmes (Catia, Solidworks), Siemens (NX, Solid Edge) and PTC (Pro/Engineer, Creo, Onshape). The management at Autodesk believed that whoever would be first to deliver on the next generational change point, from desktop to cloud, would be the new dominant player.
Autodesk started to develop a new cloud-capable manufacturing tool which became Fusion 360. With that work, Autodesk experimented and formulated how a cloudbased application with a thin desktop client should be architected.
Anagnost commented on this synergy, “Forge, and its relationship with Fusion, is quite deep and quite intense and there’s more Forge integrations with Fusion than there are with any other stack inside the company right now. Forge is playing a critical role in enabling simultaneous collaboration across various capabilities and providing fluidity of the Fusion data across various disciplines from design to manufacturing, into electrical design and all the things associated with that.
“Forge is going to play an increasingly more critical role in that, across everything that our customers do, and across all the application stacks. You’ve heard us say, over and over again, more and more of those stacks are going to start consolidating in a way that looks much more like Fusion. It allows us to not only simplify what we present to our customers but ensure that we’re leveraging the full power of the cloud to the customers’ benefit ultimately.”
Fusion has been an important learning step for Autodesk as a proof of concept for its cloud-based aspirations. It was at this point that Autodesk started to make some foundational decisions which are reaching maturity almost 10 years since it started. The company embarked on a development plan to turn all its discrete software components into an array of cloud services with customers’ data at the centre, in the cloud.
If you want a bit of a looking glass into how Autodesk’s Forge applications will act when replacing the desktop applications, it’s worth getting a demonstration of Fusion.