For several years we've heard that SIP as a powerful tool can change what we can do with business telephony. I bring this up because as an installing company I've noticed some key differences about SIP over TDM/proprietary solutions. Which of these if any, will I or other installation companies or users object to, eventually overcome or simply dismiss? Are these barriers to slow down adoption and purchase of SIP solutions, or are they anticipated behavior changes in how technology works moving forward? Preparing the users or customers means setting expectations.
First thing notable about a SIP telephone is the length of time it takes for that device to come online. My tests are simple observations and remember there are numerous products and customer networks and configurations that will vary. What could change the time to restore--cable length, network congestion (LAN or WAN), processing load on the IP-PBX, server or SIP server (host).
Here are some examples from our IP-PBX, and the hosted solutions I previously clocked were over 7 minutes:
So in my case, my SIP phones boot up faster than my IP phones but nowhere near as quickly as TDM-DPT. Even faster than DPTs are the 2500 sets--plug them in and pickup the handset and dial tone is there. Some may say about the lag times that this is a trade off or it's a deal stopper while others may not know what to say until they go to use their phone and need a reboot.
This may seem trivial but I've been observing another small detail about SIP telephones. Depending upon whether or not you are using the factory SIP branded phones or an OEM phone, your telephone clock is either getting timing from the IP-PBX or the phone is getting it from an external NTP server. How much more traffic does it generate? Small installs it's probably negligible but for those of you in large enterprise I'd suggest you start counting packets.
In the case of SIP trunks over analog trunks and in my case (my gear), it takes 5 minutes for the SIP trunks to reconnect from my IP-PBX to the host provider's SIP server unless we are aware of the failure, then we can log in to the IP-PBX and reboot the virtual SIP chassis without rebooting the entire system. Analog CO lines, as soon as you plug them in, deliver near instantaneous dial tone unless the IP-PBX has a voltage sensor with a delay timer on it before restoring dial tone. DSL and T1s don't take 5 minutes to reboot or synch up but then that depends too. On the telephony system--each IP system is different and there is a boot order of what comes up first, second, third... and each type of device requires an amount of time to get online--these are the times you need to get familiar with along with the boot order.
On the positive side of IPT, I've written in the past about answer and disconnect supervision, time to ring the device (audible to user) and the benefits. For example, incoming calls to a POTS service through an IP-PBX ring about 1.4 seconds faster (in our test case) on IP telephone sets than TDM proprietary sets. Time and timing is everything. Look to IPT for large installations and the benefits add up on a per call basis meaning less trunk and IP-PBX occupancy. When you use IP with digital trunks the same is true--they are generally faster with IPT systems and the time benefits should not be ignored.
It isn't easy getting folks to understand that devices booting up beyond a couple of seconds is an eternity. To put it into perspective, let's take a 45 station hosted platform; the local LAN switches go down for whatever reason. It will require 315 minutes total user wait time for their phones to reboot or 7 minutes per user in the case I wrote about. What is the average salary dollars for this group of users? This argument is no different from that of the financial planning process for executives electing to purchase, charter or co-op payments for private jets.
I want to add some emphasis on timing and the benefits of reducing call setup time, call disconnect or tear down, ring time, time in queue or anything to do with time and telephony including the above examples. I hope this captures your imagination--the point is, time costs money. It may seem insignificant to you, but it's not. The more time you add to a "system" the more hardware, processing power, resources and bandwidth you will need. Blocking, non-blocking, virtually non-blocking are former telephony terms often ignored by the newbies in telephony. They all relate to time. Shave off 1 or 2 seconds on every phone call for a sizeable call center and you are doing something very positive. IPT has the potential to do this, but it's a double-edged sword--it also has the potential to rob you of time and lots of it.
For those of you involved in new cabling projects: design your cable plants around "centralized" IDF locations and find ways to reduce the length of your horizontal copper drops and distances you run your copper riser connections. You are adding too much latency by going by the "standard" of 100 meters or 330 feet and you add to the cost of cabling.
Each IPT system is going to behave a certain way and it's really up to each user company to understand the timing differences and behaviors of their gear. Knowing up front what to expect could either trigger some concerns for improvement or validate some benefits for proceeding to adopt the technology. Just remember--timing is everything and getting everyone's expectations onboard at the same time isn't anything less.
