Combining mobile technology and cloud computing when using CAD/CAM
Key points in this article
- It is important that companies use cloud hosting in the right way
- Moving modules to the cloud will offer more computational power
- The application moves behind a low-bandwidth, high-latency network
- This endangers the fast-response graphics key to engineers’ productivity
Combining mobile technology and cloud computing holds promise for the CAD/CAM community, writes Peter Thorne, managing director of Cambashi.
Considering what mobile technologies have to offer the CAD/CAM community, the case for them appears to be fairly straightforward.
Connected tablets, laptops and smartphones allow users to access CAD designs, engineering applications and data, as well as interact with the project team. These devices can be used in the field, where actions and decisions can be taken immediately.
But as mobile usage among CAD/CAM users grows, so too do their data requirements and the need for storing, backing up and sharing data anytime and anywhere. Naturally, a lot of this data could move to the cloud, but even then there are still important connectivity and synchronisation issues to address.
There will be security and user-interface issues to solve, but these are largely tractable. A more critical problem arises when the connection becomes unavailable especially if you are trying to synchronise changes and edits within a project, working with a team.
This can be a slippery slope and you need to be ready to enter the complex world of ’synchronisation’, the catch-all word for getting all these user updates back into the shared database. When it comes to synchronisation, IT decision makers face some tough choices.
As mobile usage among CAD/CAM users grows, so too does their need to share data anytime and anywhere
At one end of the spectrum no updates, shared offline data is read only there are few technical problems. The mobile device can cache data automatically and the user can make copies as needed. Although you need to be careful about allowing links to the local copies, this approach allows users to create ’new’ data on their devices and add them to the team’s shared dataset when reconnecting. This approach is relatively straightforward in IT terms but can be clumsy for users.
At the other end of the spectrum unfettered offline write access to shared data you need to get ready for the sophisticated world of shared distributed databases. In turn, this means control locking, synchronisation and the multi-phase commit and rollback processes needed for the reliable handling of offline changes. Of course, there are options between the end points of this spectrum that need to be carefully considered.
Cloud computing covers quite a range of technologies, one of which is ’hosting’. Evangelists will say that this is just a tiny part of the proposition and that the real value of the cloud is in the flexibility, support for networked organisations and the vision of composite applications and datasets.
But to get there, it is important that companies use cloud hosting in the right way. The likely impact of cloud technologies on the provision of the computing, storage and I/O capabilities needed by engineering software is largely independent of whether the cloud is public, private or hybrid; or how the technology involved maps to the interface being software, platform or infrastructure as a service.
With all the hype surrounding the cloud, serious studies provide a refreshing counterpoint
With all the hype surrounding the cloud, serious studies and reports provide a refreshing counterpoint to promotional materials. They quantify and dissect the efficiency and flexibility of the cloud’s server sharing. However, most of these reports focus on general commercial computing, such as performance benchmarking or computing in areas such as molecular analysis.
Engineering IT has elements that fit these domains, but also includes software with quite different characteristics for example, interactive CAD and CAM systems with 3D graphics or complex simulation algorithms. How do these fit the model?
Take a design engineer who uses the 3D CAD modules of an integrated PLM solution.
For the cost of the high-performance workstation, moving these modules to the cloud will offer the engineer much larger amounts of computational power, as and when needed. The workstation can be replaced with a simpler device that is smaller, cheaper and simply plugs into the network to access the remote modules.
So while the business case seems solid, a cloud-based solution may still not be ideal.
The 3D CAD application is computing intensive, but that is only part of the story. It is also capable of high I/O bandwidth. This is critical for immediate graphics response when moving around geometry, so that the new positions and shapes of connected features such as fillets can be viewed immediately.
By transferring functionality to the cloud, the application moves behind a network with low bandwidth and high latency, compared with the bus-like speeds available in the local workstation. This endangers the fast-response graphics functionality that is key to the engineer’s productivity.
The trade-off between network costs and performance is not new. In 2003, Microsoft Research’s Jim Gray wrote about the factors that must be considered to find an optimum distribution of workload between local and remote systems. He characterised the break-even point as ’a minute of computation per megabyte of network traffic’. In this context, interactive I/O for software on a remote server is, effectively, network traffic.
So while the combination of cloud computing and mobility holds a lot of promise for the CAD/CAM design community, before reaping these rewards, you have to make sure the fundamentals add up.