carolyn
08-07-2004, 04:39 AM
I was asked to review the monthly stats that are generated monthly through GMS for our endpoints. It has been a while since I reviewed the raw data it gathers and what we use for our monthly activity reports. I don't know if I stumbled across a bug or have discovered a flaw in our installation and set up of the system.
We've been using GMS since 2.6....we are at 6.0. I have 29 systems...25 FX, 3 512's and 2 VS4000's. 3 FX's and 1 512 are in IP only rooms. The rest have 384 BRI and IP. Our connections do not use speeds above 384 for either IP or ISDN. We have an MGC 100 as well...2 ISDN PRI and IP. Many of our conferences are internal MP's due to multi site locations...US only. We typically use the built in MP of an FX or VS for most MP..the MGC is used for specific hi profile VTC, for connections having more than 4 sites, connecting our IP rooms to the outside, and other specific set ups. I keep the software on all endpoints updated so they are current. We do not do any streaming, use visual concert, etc...all room pc's have their own LAN connection..basically it is a straight forward operation.
When I was reviewing the data I noticed some weird calculations. I ran the report for all the sites, IP, ISDN, IN OUT, ..etc...and reviewed that to doubdl check and saw a big issue....Furthermore, thinking it may be a bug in GMS 6.0 or FX/VS 6.0 I looked back through our reports and as far as I can tell..at least since April of 2003 (the last raw data I could find)...the same issues exist. I looked through Polycom's website and forums like this for any info and found nothing...am I the only one? Let me briefly explain it as best I can and you tell me...
Upon initial review...I noticed that GMS was registering a site's activity under the wrong connection type...It would list site A's ISDN activity ...show the IN/OUT and destination (from or to)...what was odd is that it would list the call as ISDN:172..(ip address of other site) instead of ISDN: 5551212(isdn # of other site)...this was true in both IN and OUT ISDN...and it would be recorded as a single ISDN call only (meaning the same call would not show up for site A when compared to the IP activity for site A). In looking into this wierd stat, i went to site A's web interface/call log as well as our schedule for that specific event date/time.
As it turns out, when I cross checked the sources, I discovered that it was a multpoint event with a mixed connection (using both ISDN and IP). I did the same for ther occurences reflecting this same error in the report and concluded the following:
FOR OUT listings
1. When originating EP makes a mixed MP call, GMS interpret6s the event as a SINGL OUT event for that EP. It record the event as an ISDN connection but in identifying the destination it lists the IP address of the last site it connects to in the event. Furthermore, it specifies the duration according to the length of time the ISDN site is connected to it. (if an IP site leaves the VTC connection).
2. In the EP's internal call log, it breaks down and details each site that site A has connected to in the MP event.
FOR IN listings
1. If site B was a participant in the event (A originated), and was connected as the ISDN site...GMS reports it as an ISDN connection for Site B's location (ISDN:5551212...IN....384)...which is correct. BUT
2. If site C participated via IP connection (A originated), GMS will record the connection as an ISDN connection but use the IP address of site A (ISDN: 172..a's ip...IN....384) and this is incorrect as it is an IP call and should be under site C's IP activity.
What's EVEN more BIZAAR is
1. When we use the bridge..bridge dials out to A...IP and then connects to D using ISDN...GMS will report this as an ISDN connection and list the IP of the bridge.
2. If you connect to A from a PC using NetMeeting...GMS also records it as an ISDN connection but uses the PC's IP address to identify the connection destination.
ON THE IP activity recording
1. It appears if B connects to A via IPO and to C as ISDN...GMS record as IP but uses the ISDN address of B. Basically the opposite of above.
When cross referencing..ALL EP call logs are accurate. I did see a coup-le of events whereby an IP only room fell to the recordings of GMS for ISDN. I did not see where it involved the 512 systems.
Why did take so long to uncover...we have realy begun using the mixed connections in MP from the polycoms over the past 3 months or so.. before, most MP from a polycom were all IP(and I have not been doing the data collections for a year). Nothing changed when we have upgraded, swapped out dead systems, etc...like I said..I have seen it in the data as far back as April 03.
What does this mean...well, we use this data to accurately report activity for proper planning and forecasting..and for our executive business summeries
Now that I have uncovered this..I am going to do some controlled testing. I am aslo going to Polycom with it..but I hope this is the result of something with our set up and not an inherent bug that has gone undetected by anyone till now?
I would bve intrested to see any explainations, similiarities, experiences,,,,etc.
We've been using GMS since 2.6....we are at 6.0. I have 29 systems...25 FX, 3 512's and 2 VS4000's. 3 FX's and 1 512 are in IP only rooms. The rest have 384 BRI and IP. Our connections do not use speeds above 384 for either IP or ISDN. We have an MGC 100 as well...2 ISDN PRI and IP. Many of our conferences are internal MP's due to multi site locations...US only. We typically use the built in MP of an FX or VS for most MP..the MGC is used for specific hi profile VTC, for connections having more than 4 sites, connecting our IP rooms to the outside, and other specific set ups. I keep the software on all endpoints updated so they are current. We do not do any streaming, use visual concert, etc...all room pc's have their own LAN connection..basically it is a straight forward operation.
When I was reviewing the data I noticed some weird calculations. I ran the report for all the sites, IP, ISDN, IN OUT, ..etc...and reviewed that to doubdl check and saw a big issue....Furthermore, thinking it may be a bug in GMS 6.0 or FX/VS 6.0 I looked back through our reports and as far as I can tell..at least since April of 2003 (the last raw data I could find)...the same issues exist. I looked through Polycom's website and forums like this for any info and found nothing...am I the only one? Let me briefly explain it as best I can and you tell me...
Upon initial review...I noticed that GMS was registering a site's activity under the wrong connection type...It would list site A's ISDN activity ...show the IN/OUT and destination (from or to)...what was odd is that it would list the call as ISDN:172..(ip address of other site) instead of ISDN: 5551212(isdn # of other site)...this was true in both IN and OUT ISDN...and it would be recorded as a single ISDN call only (meaning the same call would not show up for site A when compared to the IP activity for site A). In looking into this wierd stat, i went to site A's web interface/call log as well as our schedule for that specific event date/time.
As it turns out, when I cross checked the sources, I discovered that it was a multpoint event with a mixed connection (using both ISDN and IP). I did the same for ther occurences reflecting this same error in the report and concluded the following:
FOR OUT listings
1. When originating EP makes a mixed MP call, GMS interpret6s the event as a SINGL OUT event for that EP. It record the event as an ISDN connection but in identifying the destination it lists the IP address of the last site it connects to in the event. Furthermore, it specifies the duration according to the length of time the ISDN site is connected to it. (if an IP site leaves the VTC connection).
2. In the EP's internal call log, it breaks down and details each site that site A has connected to in the MP event.
FOR IN listings
1. If site B was a participant in the event (A originated), and was connected as the ISDN site...GMS reports it as an ISDN connection for Site B's location (ISDN:5551212...IN....384)...which is correct. BUT
2. If site C participated via IP connection (A originated), GMS will record the connection as an ISDN connection but use the IP address of site A (ISDN: 172..a's ip...IN....384) and this is incorrect as it is an IP call and should be under site C's IP activity.
What's EVEN more BIZAAR is
1. When we use the bridge..bridge dials out to A...IP and then connects to D using ISDN...GMS will report this as an ISDN connection and list the IP of the bridge.
2. If you connect to A from a PC using NetMeeting...GMS also records it as an ISDN connection but uses the PC's IP address to identify the connection destination.
ON THE IP activity recording
1. It appears if B connects to A via IPO and to C as ISDN...GMS record as IP but uses the ISDN address of B. Basically the opposite of above.
When cross referencing..ALL EP call logs are accurate. I did see a coup-le of events whereby an IP only room fell to the recordings of GMS for ISDN. I did not see where it involved the 512 systems.
Why did take so long to uncover...we have realy begun using the mixed connections in MP from the polycoms over the past 3 months or so.. before, most MP from a polycom were all IP(and I have not been doing the data collections for a year). Nothing changed when we have upgraded, swapped out dead systems, etc...like I said..I have seen it in the data as far back as April 03.
What does this mean...well, we use this data to accurately report activity for proper planning and forecasting..and for our executive business summeries
Now that I have uncovered this..I am going to do some controlled testing. I am aslo going to Polycom with it..but I hope this is the result of something with our set up and not an inherent bug that has gone undetected by anyone till now?
I would bve intrested to see any explainations, similiarities, experiences,,,,etc.