Last May I wrote an editorial entitled “PLM and cPDm Update”. That editorial appeared soon after Dassault Systemes announced in March 2006 that it would acquire MatrixOne for around $408 million. For its most recent fiscal year ended July 2, 2005, MatrixOne reported total revenues of $124.1 million. MatrixOne was folded under the ENOVIA brand name along with ENOVIA and SmarTeam. MatrixOne itself had acquired Synchronicity Software, Inc. for consideration valued at approximately $18 million in stock and cash in June of 2004. At the time Synchronicity, founded in 1996, had more than 120 electronics industry customers, including 13 of the top 15 semiconductor companies. The company was private and had raised $31.5 million. In addition to venture capital investors Intel Capital, Cadence Design Systems and Synopsys were strategic investors. In its last fiscal year Synchronicity generated $12 million in revenue and had a user base of around 25,000 seats. At the time Mark O'Connell, then President and Chief Executive Officer of MatrixOne, said “Synchronicity will add compelling new solutions for semiconductor and electronics design management to our solutions portfolio, expand our customer base by approximately 20% and deepen our domain expertise in the electronics industry.” On February 20th Dassault announced the release of three new products for the semiconductor industry: Enovia MatrixOne's Synchronicity DesignSync 5.0, MatrixOne Semiconductor Accelerator for IP Management and the MatrixOne Semiconductor Accelerator for Team Collaboration.
I had an opportunity to discuss the situation recently with John Stanglewicz of ENOVIA MatrixOne Synchronicity.
Would you provide us a brief biography?
I am the Director of Semiconductor Products for Enovia MatrixOne. I come out of the Synchronicity acquisition a couple of years ago. Responsibility-wise this means I am responsible for engineering product management, both the MatrixOne products and the products we kind of affectionately refer to as the former Synchronicity products as we well as how these products are being merged, how they are being combined and how they fit into the whole PLM structure and value proposition. Prior to that, I used to work at a company called I-Logix which included some embedded software work. We also had a foray into HDL generation from state machines for a while. Prior to that I was formerly a customer at a couple of different companies, Hamilton Standard and what was then McDonnell Douglas and Sanders. I originally had a military background.
Would you provide us a little history of Synchronicity up until its acquisition by MatrixOne in August 2004?
Synchronicity started about 10 years ago. The vision of the founders was that the need for design data management was very pressing in hardware design. Our original ancestry was in analog design. As you probably know data that is produced by a lot of EDA tools is fairly complex with respect to the number of files and how they need to be managed together. The vision was to be able to provide a more natural way to manage these files so that each individual user does not have to understand which files needed to be kept together. Take for instance Cadence data. You have four, five or six files that need to be kept together in order to represent the information in the cell view. To be able to version manage, to combine and to collaborate on different designs meant that anyone who was trying to create a version or keep track of the history or bring cell views together had to understand exactly the directory structure and all the details and files necessary. If you left one file behind, you could easily destroy that cell view and a lot of work. As the design rolls up to the final tape out obviously missing a single file could be quite disastrous. DesignSync is a design data management solution built to solve that problem and to provide a collaboration environment for users. As time went on they added a product called ProjectSync which extended the collaboration engine to include the ability to dialog and track details about things like issues. People often used the capability to collaborate on things like features, requirements, and even task check lists. It was a very extensible and flexible tool. It allows people to have a centralized way to say I’m going to create an issue that directly attaches to DesignSync data. Then multiple people could enter a dialog about that issue, keep track of status, walk through things like approvals and close out issues while making sure all that information was available in a central way such that people could always have that status and history. Later in the company’s history they introduced the concept of an IP reuse catalog based upon the underlying DesignSync vault. As designs reached the complete state, you would have a place to catalog and reuse design. It would be natural then to vault these designs in a DesignSync structure and provide the user a way to say “Here’s a version that is complete. Yes I still need to be able to keep track of it. I still need to be able to collaborate on it because people may download and use it. I would like to know who is using this design in case there are issues that are found downstream that need to get back to the designers. I may need to have similar collaboration discussions around this. I would like to notify other consumers of this piece of IP whether this was internally developed or externally purchased. I want people to be able to subscribe to informational broadcasts or emails about new things that are created or when updates become available.” Once we had this whole tiering of architecture, people in the company began to talk in a manner very similar to that of product lifecycle management. We were effectively providing a way to manage the lifecycle of this EDA design all the way from the beginning to work in progress through full development and publication and even into its reuse lifecycle. I think that in part this is what attracted MatrixOne to eventually purchase us. From my point of view we seem like a deep dive into the electronics and semiconductor industries. We really expose the kind of things that they were always trying to do in PDM first and then in PLM. But we at Synchronicity dive down deeper and start maybe a little bit earlier and have a very robust lifecycle around work in progress.
How big a company was Synchronicity when MatrixOne acquired it?
I don’t know the exact count. But it was probably near 100 people as a pre-IPO company. I don’t know what the revenue numbers were.
When MatrixOne took over in August 2004, what changes were made in how you did business or in product direction? Was Synchronicity left to operate with a large degree of independence?
I would say a little bit of both. There is certainly an audience that is doing things at a different level than some of the other PLM customers in some other industries. I know one of your questions is “Why is PDM in EDA is so different from MCAD?” We have a lot of use models that get into detailed design processes. A lot, especially the DesignSync product, have their own progression. When it comes to ProjectSynch and IP Gear, the issue management and collaboration part, and the IP reuse part the MatrixOne acquisition was very natural and fortuitous for us. At the time we had been looking at how we could create a more enterprise wide version of these solutions. In part we were looking to build out these products in a Java engine, a J2EE environment. This is one of the things that the acquisition brought to us, an already made enterprise server. This is part of what we just released as the next generation of these products. In this new environment they are clearly more scalable than the original products years ago. The Java environment also means that instead of being little old Synchronicity and building our own, we are able to use something that is highly scalable and has already been deployed at many, many customers. We are able to inherit functionality that has already been produced rather than having a long list of things that we would want to eventually get on this new platform. It is very natural evolution for us in those regards.
It has also brought too us a couple of things which I did not foresee at first. There are a lot of areas of synergy. One of the areas that our customers tried to pull ProjectSynch into was broader project management. They were looking for things like work breakdown structures and task management. That was one of the first projects we created after the acquisition even prior to migrating the ProjectSynch and IP Gear material over to MatrixOne. We created a product called Semiconductor Accelerator for Enterprise Product Management. What this did was allow the DesignSync vault and work that the end user did at the design level to drive work breakdown structures in the MatrixOne platforms in a product called Program Central. The combination of the value of these two things is bigger than the sum of the parts because neither of these could really address this space without the other. DesignSync has detailed control and handoff of the data that the engineers were checking directly into Cadence and Synopsys environments. We were able to provide a natural way to drive these schedules. Otherwise people were forever going to use sneaker net for updating managers and managers of managers of the status. This sort of synergy was kind of a surprise for me. I think we will have some other areas of possible synergy. In a lot of ways we have kept to our strength but we are finding a lot of ways the larger organization the larger set of functionality are giving us ways to add things that were not in our plan years ago.
Were there any ways in which the way Synchronicity was sold that changed after the acquisition?
It was evolutionary, especially right after the acquisition. Existing sales channel really understood the customer base and the technology. It was at a much more detailed level than they were selling in the PLM space. Over time especially as we started to integrate products into the PLM framework the sale forces really blended, transferring knowledge in both directions. The detailed workflows and design flows that we knew from DesignSync are migrating into the broader salesforce so that they can support it. We see people even in other industries are beginning to ask about DesignSync. Similarly PLM, that enterprise level of management, is bleeding into our original Synchronicity salesforce who are more EDA oriented and more project oriented. Understanding enterprises and how value cuts across enterprises is very important to them. Right now we are pretty homogeneous in most directions although people still have their specialties.
In the spring of 2006 Dassault Systemes, a broad PLM provider with two existing PDM solutions (Enovia and SmarTeam), acquires MatrixOne. How does Synchronicity fit into this new picture?
I am not the right person to answer that question. I can speak from Synchronicity’s point of view. I can see that when MatrixOne acquired Synchronicity it was a very clear part of their strategy to get deeply into the semiconductor space and blend the value we had at the work in progress level with the enterprise level value that MatrixOne provided. When MatrixOne got acquired it became clear that the semiconductor segment was very important to Dassault Systemes going forward. I clearly see the logic of that. I am more hands distant from how the companies will blend and change over time
Have you seen anything at Synchronicity in terms of suggestion for changes in product strategy coming from Dassault?
It is still pretty early for the Synchronicity piece in this. One of the big things is that although DesignSync was originally focused on EDA data, we have always used it to manage our code and our products. As soon as we were acquired by MatrixOne, they began to use DesignSync to manage their code. This is our development environment, our repository, with vaulting and configuration management for software. This is beginning to bleed into Dassault. I hear some talk about broader plans to get into the larger organization. That’s definitely important to us. Clearly our customers are also asking for more software capabilities and are trying to use an environment that bridges both hardware and software. We see this as a strength for us as we go forward into the future.
Who do you see in the EDA space trying to do something similar to Synchronicity?
If you mean the Big Three or the Big Four, I don’t think they will do anything. Cadence had done so in the past and decided it wasn’t their area of expertise. Our company’s ancestors kind of intermingles. Cadence used to have a product called TDM years ago. They end-of-lifed that in favor of a relationship with Synchronicity. Synchronicity is still the favorite environment for the company and their customer base. Design data management is very hard when it comes to trying to manage these complex set of files. It is really very tricky to keep track of versions, to give people a way to tag things and get back to different points in time, to branch and merge. I do not think it is the best way for them to spend their resources. They have other very difficult problems they are trying to solve; OpenAcess and other kinds of data base interoperability. They are all stretching further into the digital space and higher up in terms of abstraction and system level design. At least I hope they don’t try something. I do not think it would be a wise thing for them to build. It is a different kind of problem which history has shown we are good at. They have decided that it wasn’t the best thing for them in the past.
Does DesignSync have tight links to Cadence, Synopsys and so forth to effectively manage data generated by their tools?
Yes, we do. In fact we create and manage some of the pieces of the environment that Cadence uses to expose their information to us. We also have tight links into Synopsys. Further, we provide a system we call custom type system that allows partners to create interfaces. You may be aware of the work that has been going on with Mentor. In that case they are doing the work themselves. The ideas is basically to give the kind of services needed to identify and keep this type of data together and decide how it should be recognized and managed as coherent objects. We do the work from our side as well as allowing customers and partners to do the work on their side. Custom type systems are also important where customers, especially larger customers, have their own internal tools like layout tools. That’s very important for them to be able to create an interface between their proprietary tools and DesignSync.
Editor: DesignSync DFII for Cadence and DesignSync MW for Synopsys
Prior to the Dassault acquisition how large a percentage of MatrixOne revenue was Synchronicity?
I think the year before we were roughly in the 25% range.
If you look at the three major high end MCAD/PLM company, each has products (Dassault – Enovia, SmarTeam, and MatrixOne; UGS – Teamcenter, acombination of iMAN and Metaphase; and PTC – Windchill) and significant revenue streams in the PDM or cPDm space. There seems to be no counterpart in EDA despite the fact that your sales collaterals make a compelling case for the need and benefit of this for EDS end user firms? Any explanation?
I have seen the question phrased as “Why are they behind?” I think that is an accurate observation. It is growing and this industry is becoming aware of this but they have simply been behind. For a long time EDA development was focused on design. The designer was almost a black robed wizard who lived in the backroom and you did not interfere with for fear you might disrupt him. You just hoped that everything came out on time. Even if it didn’t, as long as it was advanced and highly technical, the delay didn’t matter as much. Life has changed. Clearly cost management has changed things in the EDA industry. The need to hit schedules is clearly becoming much more pressing than it was as much as a decade or more ago. There is change going on. They are realizing you can’t just focus on the design team. You need to focus on the whole process, the whole development from beginning to end. They have also realized that you can’t leave the designer in the backroom to just do their jobs and hope they come out in the end with the right thing. You need to manage them through that lifecycle. The other teams need to get at that information and collaborate on it earlier. There was a cartoon in the New Yorker called a New Yorker’s view of the world. The picture was supposedly of the entire world but three-quarter of it was 4th to 6th Avenue. There was a slice that was New Jersey and there was a dot on the horizon that was CA. There was another smaller dot that was Japan. This used to be the electronic companies view of producing a semiconductor. It was all design. I know there are schedules and other things going on and I know what goes into a carrier and I’ve got to worry about bonding eventually but all that is pretty easy compared to design. This view is changing and that horizon is coming closer. Certainly, they now realize that collaboration is necessary. It is a bigger part of things. I have got to connect these guys to the rest of the enterprise. I have to be able to get cost information. People are talking about things like package centric design. How do I make sure I am taking into account the package goals? You can not treat this as a self contained block any more. Wherever we are going to end up in this quest for design for manufacturing, it clearly means more interchange than there was before. Why it took a little longer for EDA to reach this stage of maturity is harder for me to guess. But my observation is that we are just in a different place.
In the next couple of years we will start to see and we certainly see it now with our customers a lot of companies that did not have an enterprise wide view of the process are now coming to us asking “How do I connect to my manufacturing? How do I get global project schedules that cut across all the different blocks in my IC and against the manufacturing schedule as well? I need an enterprise wide view.” I think this in an evolution whose time has come. In the next couple of years you will really see PLM and PDM hit mainstream in the semiconductor environment.
One of my explanations is that a car or plane contains thousands of components that are independently designed, manufactured and often purchased from third parties. In the case of a chip there are millions of transistors but they do not come from independent suppliers.
The complexity is at the design stage as opposed to the assembly stage.
You are beginning to see a little of that with IP. There are not thousands of different IPs on a given chip. But there is an increasing need to manage the data associated with this third party IP particularly, if the IP is to be possibly reused in a later design. The IP itself has a lifecycle with different versions.
That is kind of hard to do if you realize that each of these blocks is in itself maybe millions of files. How do I bring that block together with my block which is also going through evolution? Maybe I have a half a dozen different design teams in different parts of the world, some of them partners, some of them internal. But each of these has its own lifecycle going on that needs somehow to be managed, not as millions of files times a hundred blocks. I have to make each of these blocks a coherent SD and it has to have a lifecycle. I need to be able to tag and manage and do all the things I used to do at the file level. I need to be able to do at the module level. This is exactly what our latest DesignSync release is providing, the ability to have a module-wide lifecycle for all this data and be able to treat it as a coherent entity rather than as millions and millions of individual files.
Does a customer purchase a central server and a bunch of seat licenses?
Most of our customers buy licenses at least for DesignSync on a project basis. They tend to use the TBL model which is very popular. We have many customers who roll that up to an enterprise agreement but it tends to be installed and used on a project by project basis. I would say that the server is centralized from the project point of view because you want a single source of the truth. The software configuration management tools take a different perspective of their space, replication of the vault and forcible merging and things like that if you look at a ClearCase model. This is not appropriate for the EDA world. That being said, we find that all the semiconductor companies do not want a single server across their whole company. They want to put the server near each project so that the mass of data being produced is localized and then design teams might have another server and combine things on a module by module basis. That is how the module aspects of 5.0 bring the value together. We also provide a lot of capabilities that allow you to have that central server and have that data cashed locally. Both caching and mirroring technologies are important. Cache being pretty much what you would expect, just a local cache of individual file versions that are close to the people doing the work. A mirror is more sophisticated. It is a specific configuration, a sort of a specialty cache instead of a pile of versions. This is the gold set of data or the bronze set of data or the set of data that is ready for test or whatever configuration is important. All these things are very important, to be able to share data, to be close and to have quick access for the people doing the work. You have to check into a centralized server so that you are not going to have the designs diverge.
The website mentions semiconductor accelerators. Are these optional add-ons to a core system?
Semiconductor accelerators are the PLM side of the product. They are integrated with DesignSync but these are the enterprise applications. The Semiconductor Accelerator for Enterprise Product Management (epm) is an accelerator that is based upon MatrixOne Program Central product; project management, work breakdown structures, task management. It uses DesignSync data to drive the schedule and automatically update the schedule. The DFM accelerator is one in which the DesignSync data and final product is fit into a larger bill of material as you move forward in how you are going to manage the backend, the manufacturing end of the chip including the GDSII file that is now a piece of the larger BOM. It includes the die, the bonding, the wire and all the steps in the processes at that level. The two new ones Semiconductor Accelerator for Team Collaboration, reuse management and collaboration, uses the team central business metrics products from MatrixOne and it connects to DesignSync data. In this case it is the next generation of our former Synchronicity ProjectSynch product. This is a step towards moving into the larger scalable Java based world. The next generation of the Synchronicity product IP Gear; IP cataloging and IP reuse built on top of library central. Some of these are areas of synergy and some of these are areas of evolution. The product and the various interfaces to EDA vendors’ data are also described on the website.
Those are sold on a per seat basis. We do not sell the server. There is a server on the MatrixOne side which has a little bit different pricing.
Do you sell through a direct sales force, distributors or both depending upon geography?
Primarily direct. But there are some other channels as well. At this point DesignSync is all direct. The PLM products have other channels out there.
Who is the target end user; the designer, manager, librarian, all of the above?
All of the above, especially when you are talking about the PLM product. You are bringing in other parts of the enterprise. Certainly the DesignSync core functionality is still directed at designers, the QE, the people that are touching the EDA data. The other collaboration and accelerator users include the whole enterprise. DesignSync users are a piece of the picture. The whole set of products is directed across the entire company.
In the hard to define typical environment are there a dozen seats, tens of seats, hundred of seats?
It is more like hundreds of seats. We clearly have customers at the smaller end of things. The typical target at a minimum is whatever the number of engineers is. If you are a small company with 10 to 20 engineers, they would be the core users. The rest of the PLM value branches out from that. The big companies with thousands of engineers have that as the core. The PLM value is the entire enterprise. The EDA focus is from engineers out.
Is there a future direction you wish to share?
There are two pieces the digital space and higher levels of system abstraction. System languages are certainly a big part of our direction. The module approach to design is even more key for those kinds of design than for analog and mixed signal design. That is definitely driving our direction as we evolve. A big component of that is that our customers are asking us to manage their software, certainly the pieces that are related and adjacent to the design. I do not mean desktop software. I mean embedded software along with the chip. It is a big advantage to be able to manage the hardware and related software in a single environment. What we have is extremely strong for the hardware. Frankly some parts of the software are easier and have different challenges but managing them together is becoming a big strength for us.
What is the average seat price for the product?
It depends upon the configuration.
Who is running MatrixOne these days?
Dassault Systemes organizes itself according to brands. The one that includes MatrixOne is Enovia. That is managed by Joel Lemke. The piece that is MatrixOne specific is managed by Mike Segal. At the time of the acquisition Mark O’Connell was CEO of MatrixOne.
On the CAD side Dassault has CATIA and SolidWorks (an acquisition) product lines. Dassault has allowed SolidWorks to remain largely independent. You do not see a merging of the product lines under a single brand. On the PDM side you now have three products with different origins under a single brand.
I think that is a good thing for us as we go forward, being able to bring together all the different values form these three solutions and devise them in a coherent way. That might be a better question for Mike or Joel.
Editor: After this interview I happened to be talking with an EDA executive who had been at part of the Windchill operation at Parametric Technology, PTC. When asked why PLM had not been a big success in EDA he opined that Synchronicity was the one company that was catering to PLM for chip design. PLM is really a software problem by and large. With boards it is more a bill of material management issue. In the case of mechanical it is a different beast. What I have seen is that vendors of mechanical products have knowledge about the way in which the assembly comes together for mechanical devices. In order to be able to create value for a PLM environment you need to be able to manage a piece of that assembly and have it work in the context to rest of the assembly. People are using PLM to manage their product development process and changes on the development process. First and foremost mechanical design companies generally do not have experience in electronics. When they try to use PLM to manage electronic data, typically what they do is tar balls. They are effectively using PLM system as a glorified source and revision control system. For the value you get, I have released version 1 of the product and here is the mechanical piece and here is the electronic piece. You have a bill of material of the electronics. The design data is typically balled up in a compressed file. It would be one entity inside the PLM data. It would not give you generally the ability to do anything with it. Some of these systems will also display a bill of material but you do not have the ability to change the bom and have it reflected in the design. You do not have the context between the actual bom and the design scope itself. They are not integrated. In order to do that they have to have an engine that would interpret a schematic or an HDL hierarchy. By and large they are glorified revision control systems. People nowadays doing very sophisticated electronics have their own systems for managing their process. They are not going to go to a PLM system unless there is something catered and built specifically for them.
The top articles over the last two weeks as determined by the number of readers were:
I WANNA BE A MANAGER
Simucad Overhauls Entire EDA Licensing Model Simucad’s Real-Time Remix is an innovative token-based environment in which a number of tokens are purchased. These tokens are used instead of licenses to run any combination of Simucad's EDA design tools. When a tool is launched, it checks out a pre-defined number of tokens from the available pool. When the tool exits, the tokens are released back to the pool. All of this happens in real time without any contact with the tool vendor.
Pyxis Technology Promotes New President and CEO and Vice President of Engineering Pyxis promotes Joe Hutt, an industry veteran with 25 years of experience in EDA and DFM, to president and CEO. Hutt, an early investor in the company, was on the company's board of directors until July 2006, when he became vice president of engineering. Most recently, he served as vice president of technical sales for Magma Design Automation Inc., after a three-year tenure as vice president of engineering. Prior to Magma, Hutt was in charge of the Advanced Technology Group at Synopsys, Inc., where he managed the Synopsys research organization. Pyxis also announced that Mitch Heins will move from vice president of business development to vice president of engineering.
Global Unichip Expands Use of Magma Software, Adopts Quartz Verification Products for Taiwan's First 65-nanometer Tapeout Global Unichip Corp., a leading SoC design foundry, used Magma's Quartz DRC Litho physical verification product for the first 65nm tapeout by a Taiwan-based design company. Already a user of Magma's IC implementation software, GUC used Quartz DRC Litho to quickly and automatically identify hotspots.
Clear Shape's Variability-Aware DFM Products are Finalists for 3 Industry Awards Clear Shape Technologies, Inc announced that their variability platform products have been nominated for three EDA industry awards: the EDN Innovation of the Year Award; the EE Times Annual Creativity in Electronics (ACE) Award; and one of the top 10 semiconductor companies' Partner Award.
Altium Designer Demystifies High-Speed Design Altium Designer 6 now adds interactive net length tuning, enhanced board layer navigation and more powerful polygon area fill placement modes to its arsenal of high-speed, high-density capabilities that already includes interactive differential pair routing, impedance-controlled routing, built-in signal integrity analysis and termination matching, automatic BGA escape routing, automatic FPGA board-level pin optimization and full PCB-FPGA bi-directional design synchronization.
Other EDA New
Other IP & SoC News