Feedback & Feature Requests
We use several Citilabs products in our company. Our current model in providing the Citilabs technology to our users and clients requires that we essentially purchase physical hardware for each instance in which the software needs to run in. Because we keep hardware for 3 years, we have some machines that run less optimally than newer machines. With Intel product releases and other advances in hardware technology happening so frequently, we are unable to take advantage unless we upgrade all computers every 6 months. If Citilabs would modify their license agreement and allow for us to install the software on virtualized machines, we can more easily take advantage of upgraded hardware and expand/contract our hardware infrastructure in our own private cloud as well as utilizing companies such as Amazon, Azure, etc. This would not decrease the number of licenses we would purchase from Citilabs. This would allow us to ensure the platform in which our software is installed on is most optimal. It's a win-win.
I would like to know what hardware you would propose to use for virtualising Cube.
In addition the Intel hardware over the past 2 years has not moved on considerably to justify changing hardware every 6 months. The core single thread CPU performance has not moved on that much. The only key improvements are the number of Cores on a single Chip.
I have seen model run times for early I7 six core chip compared to the latest Xeon E5-2667v2 processors and overall there is a reduction in run time. Unless your model takes days to run you will not be advantageous to upgrade. Typically a 10% reduction in run time. So if you model takes 4 hours to run a likely saving would be 24 minutes. in my view it would not justify the investment and migration to new hardware every 6 months.
How is your current model setup?
- Do you use Cluster
- Full demand model
- Model Choice
- Typical run time
- do you use GDB/MDB
The Cube Cloud Service is Hosted by Amazon, and for small consultants could prove to be cost effective as you only pay for the CPU time used, so no direct investment in the hardware. The model can be developed on a low spec machine and then uploaded to Cube cloud.
Thanks for the feedback. Citilabs does offer licensing options that may offer the features that you are looking for through corporate or enterprise licensing. Also there are a number of cloud options as well through our Cube Cloud Services.
Please feel free to get in touch with me about what options are available for your organization.
We have an Enterprise License Agreement that removes the need for a dongle and makes virtualization much easier to implement. We have been looking into moving to a virtual environment for over a year but have had some delays as we changed IT service vendors (more associated with our corporate network). We have been looking for examples of how people implement the virtualization. Which platform they use, etc. If anyone has a working example we would love to get the details.
I have tried contacting you several times on linkedin. We have successfully deployed Cube on a large scale. I tried to contact you to discuss what information you need.
I haven't gotten any requests to connect from you or messages via LinkedIn. However, I just sent you a request to connect. You can send me an email at email@example.com or firstname.lastname@example.org. We have just started looking into how to set up the virtual environment again. Our main issue right now is getting the IT consultants to scale the solution to something that doesn't coast $50k plus to implement. JAB