Each period in history brings its own power. Back in the 17th century, during Dutch Golden Age, the value of tulip bulbs reached extraordinary levels. Energy resources and oil were and still are some of the most powerful resources. Back in 2017, the Economist published the article titled “the world’s most valuable resource is no longer oil, but data“. Since that publication, the topic of the power of data generated a lot of interest. While there are so many arguments about it, there is no doubt that there are many ways in which data indeed can (and does) improve the world.
Data in the 21st century is an immensely untapped valuable resource. Like in other businesses, those companies that know how to realize the value of data and know-how to extract and use the data will be hugely rewarded.
The situation is extremely interesting in the manufacturing and construction industries. Industrial companies are sitting on the goldmine of data with very few options on how to turn the data into value and competitive business advantage. Therefore all discussions about how to get the value out of the data assets always trigger a lot of interest.
One of my favorite topics is graphs (or networks). The graphs are interesting and fascinating. One of the reasons is their ability to become a model to extract the value of data by building connections and networks between different elements of the information. In my last few articles, I discussed how Graphs and Network-based solutions can impact PLM and a broader scope of engineering and manufacturing software. If you missed the articles, please check them out.
In my article today, I want to discuss trends and possible options to include graphs and networks in the data architectures of engineering and manufacturing software. Here are 3 options of solution architecture that can leverage the power of graphs in existing and future solutions.
1- Linked Data Layer and Graph-Based Solutions
This option is natural and elegant. Let’s leave all data intact. The existing data is complex and hard. Call them enterprise or authoring systems and leave them to manage the data in the way they do it now. Similar to Google, which has been indexing the entire internet for the last almost 25 years, we can collect data from multiple systems and produce intelligence.
The idea of a linked data layer is powerful. I can see quite a number of companies following this recipe. I did it back in my previous company before joining Autodesk. You can find data infrastructure platforms, vertical platforms, and the IT department of industrial companies using these solutions. Here are a few examples. Neo4j Graph Databases besides being the database and providing infrastructure is promoting and popularizing the graph-based solution. In a nutshell, Neo4j is selling its database as an engine to the IT department to develop its own solutions. There are few other examples of vertical solutions, also collecting information from data sources and building solutions. One of the companies is ConWeaver GmbH, the outfit focusing on building graphs of data and developing vertical solutions.
2- Graph Database Storage
A completely opposite approach is to use Graph Database (eg. Neo4j) as a database engine for future PLM solutions. According to this group of people, Graph Database provides a much better foundation for PLM and AEC solutions, and therefore using Graph databases will bring these systems to a completely different level.
One of the examples of this approach is Ganister PLM, a new PLM company from France, which is using the Neo4j database engine as a central database to store all information. The advantages of this approach are also clear – let’s use power analytics at the core of the data management solution. The con of this solution is that although Graph Databases are powerful, they are still not a universal data management platform for PLM systems compared to other mature SQL and NoSQL platforms.
3- Network-Based Platforms and SaaS architecture
SaaS architectures were around for the last two decades and they matured into stable and very scalable systems. The architecture of SaaS systems allows the use of multiple IaaS platforms such as AWS, Azure, GCP, and others as well as to be used for private hosted platforms. These architectures apply the principles of polyglot persistence and use multiple databases (SQL, NoSQL and Graph DBs) to provide the best fit for data and application needs. SaaS architectures are growing and there are several examples of usage of SaaS architecture in engineering and manufacturing software – Autodesk Forge, OpenBOM, PTC Atlas, and some others.
The advantage of SaaS architectures is to be able to manage data as well as to use the data for analytics and graph applications. Combined with the microservice architectures, these platforms represent the online source of information that can be open and available for other consumers. Combined with multi-tenant data models, these solutions can get a unique option to analyze the data from multiple companies and get unique insight similar to what we can see in global web platforms such as Facebook, Google, and others. It is, of course, depends on the business models, security access, and many other factors. The con, if you will, of these solutions is that all of them are representing a long-term transformation from the existing architectures. However, openness is the key and these architectures can be efficiently integrated with other legacy databases, desktop, and enterprise solutions. Here are few examples from OpenBOM CAD integrations or Autodesk Forge Design Automation APIs.
Each of these solutions represents its pros and cons. They are not mutually exclusive and I can see how these options can co-exist in the future to multiply the value. At the same time, each of these options gives you a unique way to leverage the power of the data and turn it into a solution for industrial companies.
What is my conclusion?
We are going to see an increased interest in any solution that is capable of extracting value from the data. While it sounds like a simple and powerful idea, there are always a lot of difficulties in bringing new solutions to customers. The technology is not enough. Remember how Google was struggling to bring the business model and monetization mechanism to their Google Search Engine back at the beginning of the 2000s? The same is happening today with all projects attempting to bring value out of the data. These processes are built on a few fundamentals – (1) data; (2) processing speed; and (3) business solution and value proposition of the data. The company that will be able to bring all three elements together might rule the industrial world in a similar way monster like Google, Amazon, Apple, Facebook and some others are doing today. Just my thoughts…
Disclaimer: I’m co-founder and CEO of OpenBOM developing a digital network-based platform that manages product data and connects manufacturers, construction companies, and their supply chain networks. My opinion can be unintentionally biased.