Meet Actify: Making product data intelligent

actify_logo-references

Meet Actify: Making product data intelligent

Many of today’s manufacturing companies store their data in siloed line-of-business systems, like PLM, ERP, and Excel spreadsheets. The challenge with this is not everyone in the company has access to these systems, and the user interface can be cumbersome.

For more than 20 years, Actify has solved this challenge by developing products that enable cross-department collaboration, data management, and quick access to extended product data. Leading manufacturing organizations such as Plantronics, CERN, and Magna rely on Actify’s solutions to make product data available throughout their organizations, helping them make critical business decisions faster.

The Challenge: Adding features without increasing architectural complexity

One of Actify’s key products is Centro, a data management and discovery platform that securely delivers a unified view of a company’s product data. Previous versions of the product were built on SQL Server, but in an upcoming version, Actify wanted to develop more sophisticated features driven by the hierarchical nature of CAD data and product structures.

Furthermore, Actify wanted the new version of Centro to have a core system that could stand on its own as a product data management application, but also serve as a platform for tailored analytics applications and business tools. They started having a history of customer engagements where they implemented customized, one-off applications for things like engineering project status tracking, change management, and product data management, and really wanted the next version of Centro to allow them to better scale through their software versus engineering time.

The challenge for the engineering team was to implement these new features without compromising performance, reliability, and — perhaps most importantly — simplicity and ease-of-deployment. “Centro is implemented as an on-premise application, so keeping architectural complexities to a minimum is very important,” shares Ross Mills, engineering manager at Actify. “Our customers want to avoid hiring an army of IT consultants to install and configure their enterprise software.”

The Solution: One database, multiple data models

Actify’s customers are dealing with diverse data. CAD files and related documents fit nicely in a schemaless document database, but the various components of a given manufacturing product naturally form a directed acyclic graph.

“At first we thought about using separate data stores, and evaluated the usual suspects for each data model – MongoDB for the document part, and Neo4j and Sparksee for the graph part,” explains Mills. “We quickly learned this would increase our application footprint and make deployments more complicated, not to mention increase complexity in our application by having to keep data consistent across multiple databases.”

Actify had come across v1.X of ArangoDB a few years earlier, and found the feature set compelling. By tying together graph and document data models in a single database, Actify now had a reason to give ArangoDB a spin.

After evaluating multiple databases and configurations, Actify decided to go with ArangoDB due to its ability to minimize complexity while also supporting the different data models they needed. Its built-in REST interface made ArangoDB easy-to-use and deploy, and Actify was also impressed by ArangoDB’s expressive data-modeling capability and powerful query language, performance, and distributed features such as asynchronous replication.

“ArangoDB’s multi-model approach provided the answer to addressing both ease-of-deployment and the different data storage paradigms we need,” said Mills. “The AQL query language is easy to learn, somewhat familiar to those accustomed to SQL queries, and in my experience is more performant than Gremlin graph traversal queries.”

The Implementation: A directed acyclic graph to model product structures and assemblies

Actify uses ArangoDB to store product structures and their components in Centro, its data discovery platform.

A key feature of Centro is the ability to import product structures directly from CAD files, and modelling product structures that have shared components. This structure of product assemblies with shared components is represented as a directed acyclic graph of products.

Actify models product assemblies through an acyclic graph.

How Actify models product assemblies through an acyclic graph.

How product assemblies are visualized in Centro using a directed acyclic graph in ArangoDB on the backend.

How product assemblies are visualized in Centro using a directed acyclic graph in ArangoDB on the backend.

How Actify’s SpinFire Web uses ArangoDB’s graph

How Actify’s SpinFire Web uses ArangoDB’s graph and document database features to integrate data from disparate data sources.

In addition to modelling the current product graph using ArangoDB’s graph database capabilities, Actify also tracks the history of changes throughout the product graph — giving its customers a complete history of product structures throughout the database.

By combining a graph traversal with 3D search, Actify also built a “where-used” analysis to find similar components based on geometric shape. This feature helps Actify’s customers consolidate duplicate parts and inform their decisions around which suppliers they should use.

The Results: Scaling through software

Actify will continue to take advantage of the flexibility of ArangoDB as it develops new features, including Materials Library, Collaboration, and Analytics applications. In particular, they look forward to exploring how ArangoSearch will help expand their search capabilities.

“By supporting multiple data models with a single query language, ArangoDB allows us to scale through our software, not engineering time,” Mills concludes.

Do you like ArangoDB?
icon-githubStar this project on GitHub.
close-link