Will CAD / PLM companies develop DB management systems?

Will CAD / PLM companies develop DB management systems?

cad-plm-database-future

The history of databases and database management systems knows many periods. Most of us developed strong association between database management systems and relational databases (RDBMS). Since 1970s, when Edgar Codd invented relational database during his work at IBM, RDBS became widely adopted. For the last 20-30 years, most of PDM / PLM developers selected RDBMS as a default choice and foundation of their solutions. Until very recently, RBDMS was a straightforward decision to manage product data in organizations. There are multiple reasons for that. One of them is compliance to the requirements of enterprise IT. Because of wide adoption of RDBMS, enterprise organizations and their IT felt comfortable to run enterprise applications on top of well known and widely adopted database back end system.

However, changes are coming to database industry too. Database and data management technology is going through cambrian explosion of different options and flavors. It is a result of massive amount of development coming from open source, web and other places. Few months ago, I spoke at TechSoft3D tech talks in Boston about PLM and future of data management in 21st century. You can find my post and link to the presentation here. The technology shifts towards usage of different database management systems or even multiple database management solutions these days. Database is a tool. When you develop complex data management solution, you may decide to have multiple databases to achieve your goal and optimize your solution. Here is a slide from my last year presentation showing pros and cons of different data management solutions.

PLM-and-database-options

Modern CAD / PLM applications are bringing new requirements and needs to database management systems. The complexity of application is skyrocketing. At the same time, customers’ demand to have systems highly flexible and configurable. Customization, high scale and transparency, new complexity of meta data management – this is only short list of challenges PLM developers are facing these days. New cloud software development paradigms created a completely new set of needs towards high availability, performance and cost. I’ve been talking about future of database technologies for CAD/PLM solutions two years ago. Navigate here to read my previous post. My conclusion back that time that PLM vendors will face the need to have new types of databases and data management solutions.

My attention was caught by GigaOM article two days ago – CAD giant Dassault leads new $14.2M investment in NuoDB. NuoDB is four years old database startup out of Cambridge, MA is promising to shift what we know about RDBMS into the next level. According to publications and press releases, NuoDB wants to re-write the rules for a 21st century database. Another GigaOM article presents 12 rules NuoDB defines to create a superset of existing RDBMS. Some interesting buzzwords you can see there – elastic scale out, single logical database, dynamic multi-tenancy, non-stop availability, etc.

Another one year old publication informs that Dassault Systems has tested NuoDB on their desktop and web applications. This fact confirms that Dassault is looking how to future develop cloud capabilities of their 3DEXPERIENCE (also known as V6 cloud platform). Here is my favorite passage:

Dassault Systemes, maker of 3D design software, has tested NuoDB on its Web and desktop applications extensively over the past year, said CTO Dave Tewksbary. In a presentation, Tewksbary went through NuoDB’s list of feature and performance claims and gave the company strong passing grades. However, Dassault focused on smaller implementations of the database during its tests, according to Tewksbary. “We haven’t gone as far as we’d like to go.” Given the complexity of Dassault’s products, officials are eager to see how NuoDB performs when running on hundreds or thousands of nodes.

What is my conclusion? Cloud system development can change what we knew about databases and database management systems. The challenges of CAD and PLM companies to develop and maintain a diverse and highly scalable database management solution can turn them to be more interested in specialized cloud-oriented databases. Running from private and public cloud, enterprise software is not limited anymore to mainstream RDBMS prescribed by company IT. Even more, companies will be interested to develop and use the most efficient and cost effective database management solutions. It will help to develop future differentiation in technology and cost. Just my thoughts…

Best, Oleg

Share

Share This Post

  • Michael Wm. Denis

    Oleg, thought provoking note.

    When I contemplate the types of dBs – I normally start with the end in mind – is the persistence layer designed for transactional optimization or knowledge / analysis / discovery optimization?

    To me – CAD/CAM/PDM/PLM is knowledge oriented. PTC’s acquisitions leading to their SLM division is definitely knowledge oriented. And I found your article on Dassault investing in NouDB very interesting from a – where are they going with this – perspective.

    In the content (document, web, picture, graphics, video, …) world, component content management has embraced XML enabled RDBMSs (such as DB2XML, SQLServer, Oracle, PostgreSQL) and/or native XML dB (NXD) (such as MarkLogic, EMC XDB / x-Hive, eXist, MongoDB, Oracle’s Berkley DBXML, Sonic, TXTML, …).

    Physical storage of content in an NXD is a file with a defined schema. NXDs uses xQuery search and supported for large distributed computing by Hadoop, Hive, HiveQL … variants of which are used by Facebook, Amazon, Netflix and google.

    The other knowledge base is the big data / data warehouse (normally with analytics tools built on). Vendors here include Teradata, Oracle Exadata IBM Netezza, EMC Greenplum, SAP HANA, …. Being in the aviation / aerospace industry – Teradata is a mainstay of most of the companies I work with.
    So how do you see these purpose built dBs fitting into your cloud dB future?

  • beyondplm

    Michael, thanks for your commentary and questions! In my view, it is all about how to support demanded complexity, cost and efficiency. 15 years ago, the choice of mainstream RDBMS was the only possible way to bring PLM solution to manufacturing companies. Internet becomes a platform and, as a result of this, we will see more “internet born” solutions and open source based systems. I expect enterprise software companies to develop alternative ways to manage data. PLM complexity is growing and design (read “CAD / CAE”) data is moving towards the cloud. There is no out-of-the-box database system to manage this level of complexity and we will see more “database-oriented” innovation in the future. Just my opinion, of course. Best, Oleg

  • Michael Wm. Denis

    You know the two camps on enterprise systems – correction two religions as these are deeply held beliefs.

    ERP does all – bolt on where you have to vs. Best of Breed (BoB) is best + core ERP.

    The complexities of integration and data integrity in the 90s (arguably even today) motivated many CIO / IT departments into a belief that ERP solved many more problems than it created. So big packaged software vendors played on that belief to sell a lot of software – which didn’t always result in business value or strategically business agility.

    Same can be said for the PLM vendors and through organic development – or more often – acquisition – they broadened their functionality from CAx to PDM/PLM to now SLM. This definitely creates overlap with ERP vendors in the SCM, SRM, CRM and SLM domains.

    If you place multi-echelon service parts optimization, warranty management, performance based contracting (PBH & PBL) and Aircraft/Asset Health Management & Prognostics (AHM/PHM) into the SLM knowledge level capabilities then there is even more “functional” overlap with transactional ERP functionality – BUT (big BUT here) – mandates fundamentally different data & content stores and structures.

    The BoB tribe / belief has driven the need for advancements in middle ware – EAI and ESB (TIBCO, RedHat Jboss, IBM Websphere, etc…). This model looked to optimize BoBs integrated via publish and subscribe middle ware with data & content transformations for each subscriber – in order to enable strategic IT agility (plug and play) that enabled business agility.

    In the A&D OEM world – just consider the fundamental business change over the past 20 years. These OEMs revenues and profits were based on selling the assets and aftermarket parts – now the asset (say a multi-million dollar engine) is only the razor to lock in profits from the blades (financing, performance logistics & maintenance, engineering program management, …) all Service Lifecycle functions.

    Given three of the four largest aerospace engine OEMs are our clients – this is a problem and solution we’ve been intimately involved with.

    Standards have helped (ISO 10303, SPEC2000/ASD S2000M, SPEC2200/ASD S1000D and the promise of ASD 5000F for SLM to PLM feedback of CBM/AHM/PHM and warranty/contract compliance).

    But the only problem with standards is – companies have to actually implement them somewhere. If your big ERP system dB isn’t industry standards compliant then guess what … GIGO. SAP’s Netweaver and Master Data Management module prove that referential integrity and meta-data management does not exist “within” ERPs.

    Multi-tenant cloud based solutions offer the promise of addressing this challenge – IMHO – if they support BoB dB where the dB is purpose built with a cloud ESB publish & subscribe B2B mindset (maybe the term should be CSB Cloud Services Bus). And by dB I mean any persistence (storage) layer.

    A lot of what my firm has been doing over the past 20 years is taking content (optimized for internal consumption) from one company and make it usable/consumable in different ways (human consumption and IT consumption) by other companies .

    This has forced us to manage data and content in a much more dynamic way – in order to support “contextual” search, transformation, delivery and multi-purpose consumption.

    So I agree that there is currently no “out-of-the-box” dB to manage this level of complexity and there will definitely be more persistence layer innovation in the future.

    But I’m not sure the holy grail rests in any perfect dB – rather – it rests in the transport layers where context specific search, extraction/retrieval, transformation, distribution and consumption (by multiple different context specific consumers) resides.

    EMC and Redhat are doing some very interesting things in “middle earth”.

  • beyondplm

    Michael, your example and insight are very interesting. Couple of thoughts…

    1- EAI, ESB, etc. — the failure (or extreme cost) of these solutions turned lots of what you call BoB solutions into dream. On top of that, single vendor suites look very attractive. ERP, PLM and other specialized software vendors, including large service partners are succeeded to sell lots of “single platform solutions”.

    2- Data management layer for product development, manufacturing, support and service is a very complicated (maybe level up from A&D OEMs supported today by major PLM vendors). To support it, next level of data management solutions needed. IMHO, this is a reason DS and maybe others are thinking how to manage data in a more efficient way.

    3- I’m not very familiar with what EMC and Redhat are doing. Can you give me some links? I’d love to take a look.

    Again, thanks a lot for your contribution to the discussion on the blog. Appreciate that. Best, Oleg