DPAR MNO166585

[MitelInMyBlood]

There's a current DPAR on the 2K regarding calls occasionally failing between two SX2K systems in high traffic conditions. However, the DPAR lists no workaround.

We've seen the exact problem in the 2K for years, going back to at least L-stream or earlier.

In our case the problem turned out to be simple GLARE.

Granted it shouldn't have been happening with trunks available even in high traffic, but it did. Setting one end to terminal in advancing order while setting the opposite end in terminal reverse order SHOULD ALSO solve it, but may not in very high traffic conditions.

Setting the hunt order to circular (at both ends) was the workaround.

 

[GX5Krocks]

The issue in question is not related to glare it is a performance issue

 

[TheMitelGuy]

What's a SX2K?? (hehe)

 

[MitelInMyBlood]

Agreed it's a performance issue. Resources are taking too much time to clear down and return to the pool. You're only just now discovering that? Youse guys need to spend more time in the field.

What's a SX2K? Oh I know you were kidding, but a DC powered control redundant 2K brings a couple things to the table that the 3300 (IMO) has yet to fulfill. Get into the life safety/life support business (Police, Fire, Hospital, etc) where their phones absolutely ("By God") must work 100% of the time and a correctly provisioned 2K will not disappoint you.

1. a redundant file system (mirrored drives are just not the same)
2. a redundant main control (yes, resiliency I know, but making 3000 lines in a hospital 'resilient' would be an expensive, not to mention daunting task, wouldn't it)
3. hard drives whose demonstrated MTBF is measured in years instead of months (Yes Mitel, the drives in your 3300 product have a reliability issue and replacement requires taking the system offline & pulling the controller out of the rack with an awfully pregnant outage)
4. a DC powered 2K can run for many hours on a centralized carrier-class DC power system (no UPS
management & maintenance issues)
5. centralized back room line power for the phones (doesn't require expensive POE Etherswitches scattered here and there or thousands of wall-warts)
6. typically 15 minutes service interruption to perform an upgrade (or the seldom-needed drive replacement)
7. no trunking gateway required
8. all phones still work when the "network" goes down
9. new, replacement or reconnected phones "BOOT UP" in 12 seconds or less (So do Cisco's IP phones, by the way)
10. OpsMan not required

There are plenty of application scenarios where the 3300 is inarguably the way to go and a couple of the PC-based apps are pretty neat. However, as far as meeting the reliability & durability of a correctly provisioned 2K in a hospital setting, the bar has been set pretty high and the 3300 "ain't there yet".

Moral of the story: The 3300 is a service shop's dream. The 2K is a customer's dream. Before your IP zealots fully M/D the 2K I'd suggest having something to replace it. Presently you don't.

 

[TheMitelGuy]

Although I love IP, I would rather be back in the TDM world. Things just worked!

 

[slapin]

I think Mitel is working on it. But few things:
1. 3300 designed as a hybrid system. So, hopefuly..
2. Resilience is a little bit confusing and hard to configure and maintain.
3. IP phone is active device and not a part of the phone system tied up with a pair of wires. It's possible to move even more business logic to the phone itself.
4. PoE switches getting cheaper.
5. Just one network to deploy.
--------------------------------------

VoIP is young, it's getting better. I remember times when it was experimental and just a toy. It was possible to use in a lab but it was not acceptable for any production use.