I am assisting another firm with their licensing setup. They are telling me that the licensing increment of 1 hour is being revised to 10 minutes.Can I get a definitive statement on this from Bentley.
Is the one hour increment still in effect or has the policy been changed?
Wow, I just heard of the changes Bentley made. This is good news.
Many of us thought the pooled licenses under the select program would act as concurrent licenses. They were not. Now they almost are. Common sense can be used now to forecast the license usage :) The new "10 minute intervals" help a lot, and there's more behind it.
Before: The usage was binned into calender hours
A User on PC A opens a session 10:02 and ends 10:08
A User on PC B 10:18 - 10:42
A User on PC C 10:35 - 10:40
A User on PC D 10:52 - 11:10
Common sense would make you believe that the peak usage was 2 licenses: PC B and C between 10:35 -10:40 overlapped with usage. But Bentley counted different. 4 different PC's within the same calender hour "10:00 - 11:00" =4 license usages! This was really hard to explain to user groups.
From now on: The usage is now binned into 10 minute slices: Minute 1-10, 11-20, and so on. Only usages filling such a bin completley will start the count! What happens to the usage:
A User on PC A opens a session 10:02 and ends 10:08 - does not count, because less than 10 minutes use. no license usage.
A User on PC B 10:18 - 10:42 - counts as usage in slices 10:21-10:30, 10:31-10:40. the first 2 minutes are ignored
A User on PC C 10:35 - 10:40 - does not count, because less than 10 minutes.
A User on PC D 10:52 - 11:10 - counts as usage in slice 11:01-11:10. the other minutes are ignored
Result: one license used!!
More surprising: Changed after DanielB corrected another time:
PC E: Usage from 7:15 - 08:10 , fills 5 slices between 7:21-08:10
PC F: Usage from 7:22 - 7:39, fills no slice completely - no usage counted more than 10 minutes usage, will be counted in the 7:30 - 7:40 slice
Result: one license used!! Two licenses used
This is a dramatic change. Bravo! Even if users don't get the idea - what Bentley counts is less than common sense would tell. (note: AFAIK ELS subscription does not change the mechanism - usage blocks a license until midnight)
Added after input Daniel B.:
PC G: Usage from 7:22 - 7:41, fills 2 slices 7:31-7:40 and 7:41-7:50
2 licenses used
Still fair comparing it to "real concurrent count"
The following diagram posted on Bentley Communities shows the information transmitted to the hosted server. It receives the same data that client machines currently send to your deployed SELECTserver. No information is hashed since the hosted server needs the user and machine names to generate usage reports just like your deployed server.
By default, client machines send usage information to the license server using the HTTP protocol which is not encrypted since deployed servers rarely are configured with an SSL certificate. However, the hosted server has one using 256-bit SHA-2 RSA encryption. So I highly recommend configuring clients to enable the "Use HTTPS (SSL)" checkbox in the License Management Tool when using the hosted server.
The hosted server is based in a Microsoft Azure data center located in the United States. Note that our static IP address (220.127.116.11) was originally reserved for use in Brazil, but Microsoft has since repurposed it for use in the United States because of the shortage of available IPv4 addresses. While major geolocation services should show the correct location, some may show a discrepancy.
The following support solution provides some additional information regarding the differences between a deployed server and Bentley's hosted server.
I have to amend what I said here because it, also, wasn't entirely accurate. Counting does not start with the first full 10-minute window as I said previously. Rather, counting will start with the second 10-minute window as long as the total duration of usage is greater than 10 minutes. To state more explicitly what we're doing:
Take the example from above to illustrate:
PC F: Usage from 7:22 - 7:39
This usage is >10 minutes, so it will count for some usage. The 7:20-7:30 window is not full, so 7:22 is rounded up to 7:30 leaving a usage record of 7:30-7:39. For the purpose of counting peak usages in a window, we're left with 1 usage counted in the 7:30-7:40 window. So, even though no 10-minute window was full, we do count usage in one of the windows because the total usage duration was greater than 10 minutes. We will not count it as usages in 2 windows, though.