Integrating with UberCloud, the online cloud community for engineers and scientists
"Our engineering case study demonstrates that the Advania Cloud together with UberCloud’s application containers are an ideal resource for our European engineering and scientific customers who want to burst into the cloud for their high-performance technical workload which often is too demanding for their in-house computers.” "
UberCloud is the online community and marketplace where engineers, scientists, and their service providers discover, try, and buy Computing Power and Software as a Service, from Cloud Resource and Software Providers around the world.
UberCloud develops high performance, low overhead, ready to run virtual images of technical computing applications, called UberCloud Containers. Built on Linux container technology, Docker, the UberCloud containers provide the following features:
- No set up: Desired technical computing application and tools are already installed
- A fully functional desktop user interface
- Browser based remote desktop access
- Automated remote monitoring at application level
- Automated license management
- Direct access to cloud storage solutions (such as Dropbox or Box.com)
- Easy file transfers via secure file copy feature
The Business Need and Proof of Concept
Many small and medium size manufacturers cannot afford to buy a powerful and expensive compute server to be able to run more complex and larger numbers of simulations which is necessary to manufacture higher quality products in shorter time. Buying a high-performance computer for the company means long procurement cycles, HPC expert knowledge to administer and operate the computer, additional expensive engineering software licenses, and high total cost of ownership.
This case study proofs that using Advania’s cloud resources together with UberCloud’s application software containers provide an excellent alternative to owning on-premise computing resources, coming with ease of software portability to the cloud, instant access to and seamless use of Advania cloud resources, and performance scalability from few to many cores, with an on-demand and pay-per-use business model.
The SolutionThe Proof of Concept (PoC) was performed over a week period in July 2015. During the PoC UberCloud containers were used to set up a technical computing environment on the Advania Platinum instances. OpenFOAM, the popular open source computational fluid dynamics toolkit was used to simulate complex blood flow through a cardiovascular medical device. Pre-processing, simulation runs, and post-processing steps were performed successfully with the UberCloud OpenFOAM container coming with a fully equipped powerful virtual desktop in the cloud and containing all the necessary software, data and tools.
|An Advania Qstack Cloud Service demo account was created and the Advania control panel was used as the primary method of infrastructure administration.
The Advania Cloud Services user interface was easy to use and responsive.
The Platinum 3X Large instance type was selected for the PoC due to its large size.
The Platinum 3x Large specifications are:
16 virtual CPU cores
61 GB RAM
20 GB disk was selected
The instance start times were around 2-4 minutes.
Instances were easily cloned and the clone instances performed as expected.
Docker Run Time and UberCloud Container SetupThe Advania instances were accessed via SSH, and the Docker run time environment was set up. This set up process took around 5 minutes and was automated down to a single command. The UberCloud OpenFOAM container was pulled from the UberCloud private registry. This process took around 10 minutes. The UberCloud OpenFOAM container was then launched on the Docker run time environment with no further configuration or set up. To allow access to the UberCloud OpenFOAM container via remote desktop VNC service, the related ports were opened through the Advania control panel as seen below in Figure 1.
Figure 1: Firewall configuration through Advania control panel
The Engineering Use Case: Medical Device Model
Figure 2: CAD model of the blood clot filter placed inside an artery. Captured using ParaView running inside an UberCloud Container on the Advania Cloud. To increase complexity of the problem blot clots were also inserted into the model (not shown above).
The model which was set up for testing is a simulation of blood flow through a cardiovascular medical device, a blot clot filter. Figure 2 shows a screenshot of the CAD model captured by accessing ParaView running inside an UberCloud container.
The CAD model of the medical device was used to generate a mesh of 5 million cells. Then the blood flow through the medical device was computed on 16 cores in parallel over 1,000 time steps using the simpleFOAM solver. The results were then post-processed using ParaView running inside an UberCloud container.
Figure 3 shows the resulting streamlines, representing the path the blood flows in the artery and at the inlet of the medical device. Figure 4 displays the velocity plot, showing the blood flow at the inlet and the outlet of the medical device.
Figure 3: Streamlines, representing the path the blood flows in the artery and at the inlet of the medical device. Captured using ParaView running inside an UberCloud Container on the Advania Cloud.
Figure 4: Velocity plot of the blood flow at the inlet and the outlet of the medical device. Captured using ParaView running inside an UberCloud Container on the Advania Cloud.
Figure 5: Full featured desktop view for remote visualization using ParaView in an UberCloud Container on the Advania Cloud. The desktop view provides usability and eliminates user training needs.The ParaView tool running inside an UberCloud Container on the Advania Cloud and accessed remotely via VNC demonstrated good performance. The end user was able to post-process the results, view 3D representations and manipulate these graphics (pan, zoom, rotate, etc.) in real time. The full featured desktop provided the entire regular feature set of ParaView (see Figure 5); there was no training required for the user to access and use this application container on the Advania Cloud.
Screen captures were generated using ParaView and the resulting image files were transferred from the UberCloud container to a local desktop using SCP to generate this report.
Monitoring and PerformanceDuring the testing phase, system utilization information was collected through two methods: the Advania dashboard and the fully automated UberCloud Container Monitoring.
Advania dashboard offers basic, easy to use monitoring of the CPU and network utilization. The report can be run for daily, weekly and monthly intervals.
The reports update frequently and reflect the utilization of the resources at summary level.
UberCloud containers are equipped with an automated monitoring feature, which sends the user an up to date snapshot of the system utilization levels, and the progress of the running simulation.
During testing the automated monitoring feature of the UberCloud containers running on Advania resources worked as expected and the testing team was able to monitor system utilization and record when test runs are complete.
This test was not intended to achieve the best performance possible; no effort was put into tuning the compute environment and gathering statistically relevant performance data. To provide a sense of the intensity of the calculation the following rough estimates are provided.
On a Platinum 3X Large instance, the SimpleFOAM solver ran on 16 cores in parallel for 1,000 time-steps of the cardiovascular device simulation in 30,000 seconds (roughly 8 hours).
The total effort (without the 8 hours simulation run time) described above to access Advania’s OpenCloud, familiarize with the environment, setting up the OpenFOAM container, testing, developing the medical application geometry, boundary conditions, starting the jobs, and doing the post-processing with ParaView, and contacting and talking to Advania Support, was as follows:
- 2 hours in setting up the test account, getting familiar with GUI, requesting increase in quotas
- 1 hour in setting up the Docker environment, getting our base container, doing a quick test
- 3 hours in setting up the medical device simulation, doing steps like meshing, running the simulations (by the way, we ran it 5 times), monitoring, opening tickets with support, etc.
In total this resulted in a person effort of 6 hours for all the activities described above.
Business Benefits and next stepsThe tests performed by UberCloud on Advania resources proved the compatibility of UberCloud’s technical container technology and Advania’s compute resources. Using the two together, a desktop like, familiar user experience, with very low overhead, can be set up effortlessly. The major business benefits which are demonstrated by this case study are:
Benefits for the end-user:- Portability: any cloud looks like the user’s workstation
- User-friendly: nothing new to learn, ease of access and use
- Control: container monitoring allows the user to control his assets in the cloud.
Benefits for the resource provider:- Getting variability into their environment. Customers want different products which is easily
Implemented with container packaging and stacking
- Low overhead resulting in high-performance
- High utilization by better use of resources
Benefits for the ISV:- Portability; software can run on a variety of different resource providers; built once, run anywhere
- Control of software usage via container based license and usage monitoring, control of user experience
- The faster the software runs the better the user experience; containers enable porting of the software
to workstations, servers, and to any cloud.