If you followed my comments on one of previous posts – Cloud PLM and IT Basic Instincts, I promised to have a conversation with Aras about PLM architecture. On Wednesday, I had a chance to visit Aras Corp. office in Andover, MA and had long conversation with Peter Schroer, CEO of Aras. The idea behind the conversation was to discuss benefits of Aras PLM architecture. Aras is aggressive in their goal to displace existing PLM systems – TeamCenter, ENOVIA and Windchill. When you think about how a company of less than 100 people can do, you would be trying to find “killer innovation” in technology and/or business. So, this conversation was about technology. Obviously, some charts presented by Peter cannot be shared in public. So, here my “collage” of the conversation to give you feeling of the atmosphere 
.
History and PLM Architecture
Aras Corp. is probably the newest PLM software created from the scratch back in 2001. Top 3 major PLM systems were created as a result of multiple mergers, acquisitions and additional development. It doesn’t say anything about what system is better. However, coming later to the game provides some clear advantages – learning from previous mistakes and use of advanced technology.
PLM functional architecture
This is my simplified view on PLM functional architecture. It can be applied to many PDM/PLM systems developed for the last two decades (actually, it can be applied to non-PLM systems too). Almost all PLM systems were built on top of RDBMS (the exception known to me was MatrixOne, which runs on object database – Objectivity, but migrated to Oracle later). A very essential layer is ORM (Object Relation Mapper) converting RDBMS into object store and providing the data model to build PDM/PLM applications. The characteristics on ORM are important and define the level of flexibility and basic functional capabilities. Business layer (API) and applications and solutions introduced functionality of the system, including core functions, integrations with desktop and enterprise systems as well as many others.
Generic PLM Functional Architecture
Almost all PDM/PLM systems passed the trajectory of development started from the so-called “toolkit” to “solutions” with many intermediate stations. This is a place where commonality ends. Different vendors put their own functional priorities as well as specific technological and implementation choices.
At the same time, I would define ORM as one of the core capabilities of every PDM/PLM system to provide an engine to implement multiple functions. In many cases, the capability of ORM defines so called “data modeling” capabilities. It is important to differentiate technological capabilities and marketing buzzwords. One of the characteristics of ORM is to use a certain level of RDBMS capabilities can be a differentiation. RDBMS (e.g. Oracel and MS SQL) is a very mature technology. DB vendors developed highly sophisticated mechanisms to support multiple data modeling capabilities. To use a right combination of RDBM data modeling can impact significantly future capabilities of PLM system.
What is my take on Aras?
In general, I don’t think Aras’ principle architecture is different from Enovia, TeamCenter and Windchill. The implementation technique can vary, but it will deliver most of the functions available in other PDM/PLM systems. On the other side, devil is in details. Coming late to the game (around 2000), Aras had a chance to learn from many mistakes made by PDM/PLM vendors during 1990s. It is an important advantage of Aras. I’ve seen lots of focus on flexibility and efficiency, which can create implementation differentiations and potentially lead to the advantages in implementation and maintenance cost. However, these things are hardly can be visible by just reviewing architecture and can be only proven by real work.
What about other players?
I can see some similarity between Aras architecture and Enovia V6, TeamCenter and Windchill. Aras claims to be lean compared to other systems, I can see some reasons why. At the same time, these claims cannot be taken in a broad range and requires detailed discussion. I found some similarity between Aras and MatrixOne (both companies were founded around I-495 in MA ). One more vendor to be mentioned here – Autodesk. We don’t have official detailed information about what is inside of Nexus PLM. It will be interesting to get inside of Nexus technology and see it (as soon as Nexus PLM will be released).
What is my conclusion? PLM architecture influences two major characteristics of the system – flexibility and cost of change. Vendors understand it. You can see different marketing terms here – dynamic modeling, insanely configurable, etc. There are no direct dependencies between how each element of architecture impact flexibilty and cost of change. At the same time, overall, it does matter. In modern PLM world, these two things will drive future success of PLM systems and vendors. Just my thoughts…
Best, Oleg
Freebie. Aras marketing wasn’t involved, and no censorship was applied to the content of this post.
An interesting addition to my yesterday post about technological options for cloud PLM. I was reading Amazon’s announcement that came earlier this month – Amazon separates servers from IP addresses. Navigate to the ZDnet UK Blog article “Amazon Separates Servers From IP Addresses to read more. Here is how this feature explained in Amazon blog:
Today we are adding additional flexibility to EC2 instances running in the Virtual Private Cloud. First, we are teasing apart the IP addresses (and important attributes associated with them) from the EC2 instances and calling the resulting entity an ENI, or Elastic Network Interface. Second, we are giving you the ability to create additional ENIs, and to attach a second ENI to an instance (again, this is within the VPC).
On the picture below you can see how Amazon explains the topology of EIN.
Cloud PLM
There are two CAD / PLM vendors officially announced their support for cloud – Autodesk and Dassault. Nexus 360 PLM is a cloud PLM coming from Autodesk later this year. Dassault ENOVIA is a flagship product sold by Dassault. Another PLM company claims their support cloud is Aras Corp. I haven’t heard any specific cloud plans from Siemens PLM and PTC. If you’re in discussion with your CAD / PLM vendor these days, you better check if product cloud configuration supports EIN.
What is my conclusion? Even the article is a bit on a technical side, I found it quite important. Cloud is moving towards having more and more “enterprise features”. This is a reaction of cloud providers on the coming demand from enterprise IT to accept cloud usage. The critical question that wasn’t answered was about the cost. Cloud attractiveness is dependent on cost structure. Cloud providers will have to charge an additional fee for enterprise features. Will it be still attractive for company IT? A big question. The architecture of cloud systems is in a very early beginning phase. You better check it carefully with your vendors and/or partner. Just my thoughts…
Best, Oleg
Cloud is hyping in 2012. Coming Facebook IPO can only supercharge the future of cloud computing. Last year Autodesk announcement about Nexus 360 cloud PLM created a confirmation that large CAD / PLM vendors will be interested to leverage the power of cloud. Two weeks before AU2011, during DSCC 2011, Dassault Systems confirmed their plan to continue development and investment in their Enovia cloud platform.
It takes time, but economic of cloud computing is too good to be ignored by CIO. At the time consumer market already embraced cloud computing via multiple store (but not only) options like Dropbox etc., CIOs are just coming to discover it. Navigate to the following Gigaom article – CIOs come around to cloud storage. Here is my favorite passage:
“The sheer volume and availability needs are pushing cloud storage to the forefront,” he said. They have to look at the economics of cloud compared to the high-cost, high-maintenance data center storage model, he said… In short, even the most risk-averse C-level information executives are coming to realize that if cloud storage isn’t in their current plan, it will be in the near future.
At the same time, I can hear voices of customers and vendors about the fact cloud computing is still confusing. So, in today’s post, I decided to put some practical technological options about how PLM (and not only) can be delivered on cloud today from the technological standpoint.
Amazon
Amazon is Amazon. Flexible, public, cloud. Period. It is a perfect virtual environment with dollar meter. You pay for what you use. Despite few outages, AWS is pretty stable and can provide you a reliable base as a platform for cloud PLM. Most of PLM vendors talking these days about cloud are exploring Amazon as a first option. Amazon also provides probably the best shortcut between existing PLM architectures and future cloud models.
Microsoft Azure
Azure is a different type of cloud animal. If you’re familiar with terminology, Azure is PaaS (opposite to AWS, which is IaaS). I can see many advantages of Azure. It is single development platform, tools, multi-language support. Another positive side of Azure is that Microsoft can much easier force developers follow specific rules that can prevent application from misbehave. The perception of vendors and developers is that Azure is closed platform. I’m not saying it is true, but this is what I think many people assume when they think about “Azure cloud”.
OpenStack
This is a very interesting option. OpenStack pretends to become “an Android of the cloud“. Open Stack achieved critical mass to become a reality. OpenStack is IaaS environment currently supported by Rackspace and NASA. Technologically, OpenStack is a combination of storage and computing library. The easiest way to start with OpenStack is to use it on Rackspace. OpenStack objective is to convert cloud into commodity, which can be beneficial for many consumers of the technology. I can see an interesting option for OpenStack and PLM. OpenStack provides a very open and economic way to establish mini-cloud centers. It can be a foundation for cloud services available for large companies having concerns about public cloud.
Cloud Databases
This is an interesting option for PLM developers. Fundamentally, PDM/PLM is all about a database today. To move PLM database on the cloud, can be an interesting option. Read my post few months ago – Will Database on the cloud supercharge PLM for Small Companies? There are few providers to be mentioned here. I’d be starting from Amazon RDS. Another option is to use databases services created by enterprise software vendors – Oracle Cloud Databases, Salesforce.com Database and few others.
Don’t Forget IBM big blue
Big Blue IBM is also going to the cloud. However, IBM is doing it differently. It called IBM Smart Cloud. You can learn here about how IBM suggests to use these services here. In a nutshell, IBM idea is to wrap whatever you have with Tivoli cloud services. IBM is attacking cloud from a software perspective and looking how to build a cloud umbrella beyond your existing data center. IBM clearly is looking how to attract “enterprise dollars” from AWS, OpenStack and Azure.
What is my conclusion? I believe we can see lots of misunderstandings with the cloud computing in a near future. CAD / PLM vendors and service providers will be able to balance in order to dance on both sides of the solutions – on premise and on the cloud. Understanding of technological options is a good foundation towards reasonable decisions about the cloud in 2012. Just my thoughts…
Best, Oleg
