The Cloud Opens the Doors to Design Exploration for Manufacturers

There have been some interesting conversations in the CAE community around the benefits and drawbacks of the cloud. One of the points I find most interesting is around the idea of “Infinite Computing”. This is a phrase that we use here at Autodesk. Some people understand the context right away, some balk at the idea and many others are likely somewhere in the middle. What’s Infinite Computing really mean to us as users? Let’s take a look.

I’ve had the opportunity to be out ahead of the market and was involved in the early days of launching a cloud connected CAE product line – Simulation 360. I’ll spare you the sales pitch and focus on some of the lessons I’ve learned along the way.

Let’s first take a look at a short list of some of the problems we are trying to solve.

  • Computation demands are on the rise
  • Limited collaboration and communciation tools
  • Limited time and resources
Saw

Computation demands are on the rise

Model sizes are increasing. The desire to account for Multiphysics is increasing. But, more importantly the pressure to optimize is a common conversation. When I probe more on why users are not leveraging simulation more to optimize their design. Often, the most common response revolves around time. Not enough time in the day to fine tune and explore more design choices. Status quo has been that good enough will have to due for now. In some markets, maybe that is ok. But, more commonly we hear that competition is fierce and pressure to deliver faster, more innovative products has never been higher. What if you could do significantly more tomorrow than you did yesterday? What if you could exponentially increase your through put? I believe that is where infinite computing comes into play. In reality, we all appreciate that even with the cloud, computing is finite. But, the idea of elastic near-infiite computing is significantly better than where we are today.

Limited collaboration and communication tools

Even the best-in-class simulation users report that they struggle to communicate or collaborate efficiently. Each company has a variety of ways they document or share the results of their simulations. Some have a formal templated system that they meticulously document their findings and formally report their findings. Others simply create quick power points that are shared via email and others simply email. The cloud does provide an opportunity for change. We are already seeing the impact of tools like dropbox and Box. Ability to access data whenever and wherever. Imagine if you were able to access simulation data in real-time. Imagine even further that you were able to share images, data with your colleagues. Now imagine that all of this information was searchable and accessible. That is what the cloud brings to the table.

Limited time and resources

All of the above are great advancements in simulation. But at the end of the day, what we are trying to achieve significantly more than what we are able to do today. We are embarking on a cultural shift where computational resources are becoming accessible. Ability to access information is available on mobile devices. The ability to compete is real. Getting back to one of the major hurdles around simulation adoption is time. If we are able to explore our designs and run these simulations in parallel. What does that do for the design cycle? What does that do for time to market? What does that do to building confidence in our designs?

We are in early days when it comes to the cloud. Many of the speculations here are educated observations and predictions. Many are being proved as we speak, many require time and validation by the user community.

TechTalk: A real-life case study about CFD in the Autodesk 360 Cloud

Autodesk entered a partnership with the uberCloud experiment over a year ago. It was an interesting opportunity to get out in front of a community of engineers, scientists and early adopters who are interested in leveraging the cloud for computation. The experiences and backgrounds of the community are vast. Some have been involved in heavy HPC applications for many years, some were design engineers that had alot of design alternatives floating around in the minds, but didn’t have the time or resources to explore the possibilities.

Ubercloud

Our team partnered with Mark Lobo at Lobo Engineering to explore a real world application of maximizing flow and pressure drop through a valve assembly. Mark was able to run over 200 simulations and take the work that typically would take a month, down to a day.

Join us next week as Mark shares his experiences of life as a CFD Engineer leveraging the cloud.

Autodesk Simulation Flex: An Evolution of the Cloud

It’s no secret, Autodesk believes in the cloud. We’ve spoken about the benefits- computation, collaboration, storage, access etc. No other team has embraced the paradigm shift as much as the Simulation Group. We have been releasing, learning, tweaking our products to leverage the cloud over the past three years. We’ve witnessed first hand the birth of a new way of thinking about computing. As with anything new there have been some ups and downs. Let’s take a look at some of the things we learned along the way.

First, we started with what is now Inventor Optimization. We wanted to tackle a problem that plagued the design community. We wanted to create a solution that was super easy, almost automatic and allowed us to test the power of computation that the cloud brought to the table. What we delivered was a purpose built plugin that sits directly inside Inventor that allows non-experts to quickly setup a parametric optimization and leverage the power of infinite computing to identify the optimum design.

Next came along Project Cumulus. We wanted to up the stakes a bit and take what we learned from the design community and extend it to the Moldflow analyst community. We picked something a bit bigger and a bit heavier in computational requirements. The team connected our Design of Experiments (DOE) technology and the power of parallel solves via the cloud to provide a unique offering that has been bottle necked historically by computation time.

Two experimental projects that served two very different communities but they both had one thing in common, the need for computation. Taking what we learned, we wanted to broaden the circle and see if we could apply it in a more general purpose form.

Enter the launch of Simulation 360. This was a first of its kind. We took the family of Simulation desktop products that we had (Moldflow, Mechanical, CFD and RSA) and connected them to the Autodesk 360 cloud. Above is a fun video we put together to show you the offering in action (“I just saw a helicopter”). The service was offered on an annual term basis with “cloud credits” that you burned whenever you used the cloud for computing. Similar to a mobile phone plan.This provided a unique business model that customers didn’t have to heavily invest in numerous licenses for Multiphysics simulations and the additional expensive hardware required to run them. You could simply subscribe and pay for what you used. This model has been widely adopted by the community and we have had well over 200,000 simulation jobs run on the Autodesk 360 servers. Although we were pleased with the results, we also learned some things along the way.

One of the biggest takeaways was that the needs of the Community varied substantially. Some were heavy users, some used it occasionally and some used the cloud for burst computing. It became clear that we were in the early stages of a paradigm shift in computing, but in order to satisfy the needs of our customers, we needed to be flexible. We spent a better part of two years traveling around the world visiting customers, observing how they used the tools. The result, we tweaked the offering and changed the name to Autodesk Simulation Flex. It has all of the similar features as before but adds the ability to solve local. We found that small quick experimental jobs, quick tests of boundary conditions were better suited for local solving and the larger, heavier lifting was better left to the cloud. This hybrid service gives you the best of both worlds, allowing you to leverage the cloud when you need it but also not restricting you from getting your job done when you are traveling or away from an internet connection.

I believe the market as well as the technology will evolve over time. Right now, we are working and living in a hybrid world. This will likely be the case for the time being. But, it seems inevitable that we will all evolve as the tech evolves.

Curious what you think?

 

 

The Key to the Cloud for CAE is Elasticity

There are a number of changes and advancements going on in CAE. No other has had more controversy and impact than the introduction of the cloud. Many new vendors have sprung onto the market to offer an attractive alternative to desktop CAE. In addition, some of the traditional CAE vendors are providing alternative offerings to leverage the power of the new paradigm that the cloud brings to the table.

NewImage

The cloud brings to the table the potential for Infinite Computing. Infinite in comparison to what we have on desktops and servers today. Infinite in the idea that the offering is elastic. We are able to use what we need when we need it. It’s more than simply a change in technology and workflow, its around the impact the business model brings to users. When we look at the number one reason that users have reported why they don’t simulate more. It almost always comes down to time. There is a finite amount of time and resources available and we tend to make due with what we have available to us. What if we removed barriers of computation, opened the doors to allowing you to explore? Automate the process of optimization?

That is what the cloud brings to the table. Yes, it can likely solve larger models. Yes, it can provide access to more physics. But, the pure power of the cloud is around elasticity. Elasticity breaks down some of the common barriers we face around Design Exploration.

In a recent survey we conducted, it was listed that engineers are running 4-5 simulations per design. When we probed how many design changes were made based on the simulation results, the answers ~2-3. What if you could do 10x more in the same amount of time. What if you could simply double your throughput?

No question that there are some barriers to adoption. But, all of the security discussions aside. Are you looking for access to more computing power? Are you content with the amount of simulations you are able to do today? If not, what are you doing about it?

How do you manage your CAE Simulation data?

Those of us that have implemented Simulation into our design process quickly feel the pain of “big data”. Whether you are running a few simple simulations or conducting intensive optimizations on large models, you quickly realize you have a large number of files on your hand and they are big!

I find that depending on the phase of the project, there are varying solutions to the problem. Some take the problem head on and attempt to store all o their data locally. This is ok, but not very scalable and limits your access to one machine.Having access to a mapped network drive is ok, but can be limiting if you are in the middle of a project and need to access or share the information. Historically, I found that I was using thumb drives or portable hard drives that i carried around so that I could have access to my data wherever and whenever I wanted. It’s by far not the greatest solution as I still had to ensure that I was “managing” the data myself. I ensured that I was deleting what I didn’t want, creating naming conventions so that I could find what I wanted and isolating data such as images and videos so that I could share them with others. It worked for a few years, but found that I quickly was carrying around more than one portable drive. Not very scalable.

TheVE2

Then things changed. Cloud based data storage solutions came onto the market. The price was reasonable and the user experience was amazing. I began using Dropbox a few years ago and began syncing all of my data to the cloud. In the early days, this seemed like a dream come true. I simply saved the data like always but now the data was accessible to me from anywhere. In addition, I began using the sharing functionality all the time. More of my colleagues starting adopting Dropbox so we had numerous shared folders and had a makeshift project management solution, minus the actual “management” aspect. We tried other services like Google Drive and Box. Each of these had added benefits over Dropbox, but had as many drawbacks. In particular, Google Drive has Google Docs embedded directly into the workflow. However, not everyone uses Google Docs. Box has some great collaboration tools, but I find that it’s not ideal for really large simulation files.

I spoke about Autodesk 360 in previous posts.. It has some syncing capabilities and some collaboration tools built in but likely needs more Simulation specific functionality. Curious, what people are using? How do you overcome the issue of access, file size, searching, sharing etc?

Are you ready for CAE Simulation in the Cloud?

There is a hot topic in the CAE Simulation space – are you ready for the cloud? Seems like a fairly straight forward question, but in fact, the answer is not so simple. I think there are a number of variables to consider. The standard questions come to mind – security, reliability, scalability, predictability and overall – does it fit my workflow? But, I challenge you to take a step back and ascess the level of pain you are experiencing in traditional offerings or consider whether or not you are an early adopter based on the value proposition.

If we try and simplify the question. What are the benefits of the cloud? What problems are vendors trying to solve with the cloud?
Here is a brain dump of ideas to consider.
  • Access to a suite of “physics” on demand, as oppose to paying a la carte
  • Access to “infinite” compute resourcing to not only solve large models, but multiple models simultaneously
  • Ability to explore optimization
  • Anywhere access to data
  • Collaboration tools
  • Mobile access
The above list just names a few of the tangible benefits of the cloud. If you found yourself nodding, contemplating than it’s likely time for you to take the next step and evaluate what’s out there.
If the above list doesn’t resonate with you, than hang tight, we will discuss more benefits in the days to come.

 

Autodesk is presenting at ISC Cloud ’13 on CAE in the Cloud

A group of the industry’s top minds in HPC will gather in Heidelberg, Germany on September 23-24th to discuss what impact the cloud will have on their industry in the near and long term. Although many will have a common interest in the performance aspect of the cloud, I suspect there are a number of questions on everyone’s mind.

  • In addition to raw computation power, what else will the cloud provide?
  • How secure is the cloud?
  • What exists today and what will exist tomorrow?
  • What are barriers to adoption?
  • If barriers are reduced, what’s possible?
I’ve been asked to present and have chosen to talk about Re-imagining CAE Simulation in the Cloud.  I am excited to meet with other vendors, thinkers and users in an effort to explore what’s possible.
Anyone else attending? Would love to grab dinner or coffee to hear your viewpoints.

The #cloud isn’t for everyone in CAE

There has been a fair amount of momentum and hype from many vendors claiming this and that about the “cloud”. The fact is the cloud has many forms and has a potential to change the way we work. Actually, I am convinced that it will have a profound impact on the way engineers work. However, the level of impact and the pace at which we feel that impact will vary greatly. Let’s take a look at a few of the areas where the cloud will impact sooner than later.

cloud icon

Raw Computation Power

I think there is no question that engineers that are doing heavy computation are constantly craving more and more power. Whether it is to run more simulations simultaneously or having the ability to run larger individual models. The list of those simulation engineers that are happy with the computation power they have is short. The cloud can offer not only scalable computing resources, but also allows for interesting “on-demand” licensing models. But, when talking to companies about their needs today and in the future, they must feel the pain today or at least be able to predict that pain in the near future or all of this talk about raw computation power is just cool or falls into the nice to have category.

Collaborative Workflows

Many of us in CAE have been touting the importance of collaboration in CAE. When presenting to a room of engineers, its hard to find anyone not nodding when asked if collaboration is important. Same head nodding goes when asked if significant improvements could be made to their existing process. There is a tendency for collaboration to be put into a big bucket that encompasses sharing files, working on models together, to sharing of results. I think the simplest form of collaboration that the cloud could help in simulation is reporting. Whether that is sharing images and screenshots of results or more elaborate reports. Today’s methods of ppt, pdf, images embedded in email are a far cry from true collaboration. These methods are difficult to archive and make searching nearly impossible. There are a number of web based technologies available that could change the way we collaborate and report our findings. Everything from searchable databases to creative ways to visualize large data interactively. But, same as above many engineers we talk to are too busy to extend beyond status quo. Seems as though independent consultants are plagued as much as anyone when it comes to collaborative workflows with their clients. Many of the tools that are coming out on the market are geared towards small independent users, rather than large corporations. Not that engineers within large corporations aren’t dealing with the same challenges. It’s just that the appetite for change in these companies is less.

Usage Cycling

Whether you are a power user or occasional user, many of the users we speak to experience some sort of ebbs and flows of usage. Often, they either outsource the overflow or in many cases, they do without. When we look at “on-demand licenses its not only for those that are craving more power. But, is perfectly aligned to offer more physics and functionality. So, you may do linear static stress and steady state thermal 90% of the time, but on occasion require transient simulation or more advanced nonlinear capabilities. The cloud offers a unique opportunity to break the old licensing models and allow users to subscribe to a service and pay as you go. But, its clear that this user model isn’t geared for everyone.

The point here is that we are in a paradigm shift when it comes to delivery of software and many engineers will have the opportunity to make different decisions about how they want to leverage new technologies. The common theme between all of the topics above is that if you aren’t currently feeling the pain or pressure to change what you are doing today. There is a good chance that you are not ready for the cloud or the cloud isn’t ready for you. Remember way back in 1995 when we said we ould never, ever run CAD on Windows, much less a laptop. Things change.

 

CAE in the Clouds

Constant buzz about this topic. Tons of speculation on where its going, pretty sure no one knows where it will end up. But I think its clear to many that the options are endless. A recent blog post, by Devon Sowell indicates that SolidWorks is getting into the game. To what extent, only time will tell. I was a bit surprised to hear that they are looking at CAD as SaaS, at least in the traditional sense.

I tend to mix cloud computing and SaaS as one in the same. But really, I see SaaS as the business side of things – how you access and pay for the software application. Cloud computing is a bit more general and simply means (to me) leveraging the web and hosted servers/sites to perform what you do (or wish you could do) locally today.

There are so many avenues that this can take on the CAD side. I suspect that as bandwidth is increasing, we can imagine that you will be able to interact, edit and create 3D models over the web. But, I am not sure from a user’s experience this is in the near future. Tons of things have to be accounted for including RAM, video performance and overall processing power for complex CAD operations. Plus, one of the selling points of SaaS is to use it on demand, “when” you need it. This has a very part-time, occasional feeling about it. Even though CAD is easier to use now than ever before, its not word processing, so there will be a need for proficiency.

But, as we know, there are many facets to the CAD world. To me, the first logical step is data collaboration – files in the sky, on demand access from anywhere that can be viewed, mocked up and shared. Reducing the need for ftp, email and time draining “quick” design reviews. PTC is just dipping their toe in the water with their Social Product Development effort. So imagine you are collaborating with another site or a vendor or a supplier and you can all have access to the exact same CAD data on demand. You can collaborate on the spot or leave text or even video messages for each other so that time zones become less and less of a burden. This access needs to be anywhere, from any device. All the data – 2d drawings, FEA/CFD results, testing data, specs, BOMS -everything must be a click away. Some might be thinking, doesn’t this already exist in PLM?? To a point, but it is still not mainstream enough to roll out to everyone. Once this is accomplished, then I think we will see the true power of cloud computing on this sector of the industry.

The next and more true to my heart is on the computation/simulation side of the world. FEA/CFD is becoming more and more mainstream. Sizes of models is increasing daily as well as the complexity of the physics being solved. This growth will not slow down and will always be bottle necked by the computing resources available. Many have tried and have been marginally successful in providing “on demand computing”. This will continue to be a theme from the simulation folks. So you can imagine that you may run CFD from time to time, but really don’t want to maintain the computing power needed. You may run conceptual models locally and leave the big jobs to the cloud. No doubt we will see more and more of this. But how we interact with this environment is key.

If I sit at my desk and setup a model and then submit all the files etc needed to run the simulation to some server in the sky, this will take time. I also will want to interact with these models in some form or another. At a minimum, I want to know the progress, check quickly if things are moving along nicely, check preliminary results and a whole list of little subtle things that I check when running locally. The beauty of a cloud model is that I want to check that from anywhere at anytime. Might seem silly, but be nice if I had an iphone app that allowed me to see what jobs I had running and have a little dashboard that is customizable to look at certain things that “I” want to look at for all my models. Maybe I’m not an iphone guy and I simply want a text message or instant message when the model is finished. Maybe I am not a mobile device guy at all and simply want to “login” on my wife’s home computer to see how things are going. To many this might be a foreign concept and many others are probably nodding their heads because they are feeling the pain.

I am closer to the simulation folks, so I could go on for days about the potential of SaaS/cloud in this space. I am interested in hearing from true CAD folks that are struggling everyday. Where can the cloud help you?

An Upfront CAE "Computing Engine" for the Common Man

The beauty of the hardware vendors making strides in technology is a great thing for all. It simply drives the market to a point where really powerful computing power is available to the common man, not just the elite early adopters.

It was only a few years ago that the idea of parallel computing was really only available to a select few. Parallel computing, of course, is the ability to leverage multiple machines connected together via a VERY high speed network connection for the purposes of doing some sort of computation FAST. Really fast.

Those of us in the Upfront CFD world or anyone in the CAE world for that matter, can always benefit from a bit more horsepower. Whether it is to help solve a really large model that can’t be solved on a standard workstation due to its size OR to reduce the amount of time a monster model takes to solve OR to speed up the process of your multiple what-if scenarios. In other words, all in an effort to SAVE TIME.

With Microsoft jumping into this world with Server 2008 HPC Edition, it now makes it possible for mainstream engineering companies to enter the world of High Performance Computing (HPC). This is clearly not a dig on Linux computing farms. I am a huge Linux fan, writing this post directly from my Linux netbook.

So why is Microsoft a game changer? Simple answer- they are familiar to people. But there is more to the story- the hardware needed is becoming mainstream. So it is really a combination of the two. Any engineer or IT professional can now setup a “cluster” or a “computing engine” and take advantage of the power in numbers.

So, here are two scenarios that you can consider when looking at hardware for your computing engine.

You can look at purchasing 2, 3 maybe 4 workstations or (equivalent servers) from Dell or HP or whoever. You will need to ensure that they have an infiniband connection to hook the two together. Fancy word for saying that each machine has a card and special cable for the processors to communicate to each other. Don’t bother trying to connect via Ethernet, you’ll use Ethernet to put them on your network so that you can communicate to them from a remote machine (laptop, workstation etc). Connecting 2 machines would require a cable, more than 2, you will need a “switch”. The more you want the more it will cost, BUT the quicker the solve, the bigger the return.

Here are a few pics of my multiple machine computing engine.

Here is the front- obviously.


Here is a look at the back. Pointing to the two infiniband connections (not ethernet) “hooking” the machines together to form the cluster. This whole configuration is in the $10k price range. This specific configuration are Dell Precisions with dual dual core processors and 8GB RAM each. The setup is fairly straightforward and is worth the effort for the cost.

But, if you have a bit deeper pockets, not super deep – but deeper. You could go with the “cluster in a box”. This is literally, unpack it and plug it in. Really nice. The “infiniband” type connections are all built into the box. In this configuration, each computer is referred to as a “blade”.


Here is a shot of the back of the box. Each one of the “slots” with a green light is the individual blade(computer). This box has 5 blades in all. The beauty is that this works as 2 clusters – one with 3 and one with 2 computers each. Essentially, the same speed/config as above.

Extremely exciting times right now for Design Engineers that are currently in or thinking about getting into the Upfront CAE space. There is massive potential for doing way more than we have ever been able to do before.

Personally, I am anxious to see where the MCAD world, MCAE space in general, finds itself leveraging various forms of cloud computing in the near future.