Ever since installing our first IP-PBX in 1996, we've found more ways to install them using existing cable plants. IP-PBXs often work better by themselves on their own cabling infrastructure than those that are converged using just one cable to the desktop. I've claimed before--years before QoS, VLANs, network monitoring and assessments and other tools were typical--that by using available resources and installing the IP-PBX as a "phone system" is better for some customers.
Ever since installing our first IP-PBX in 1996, we've found more ways to install them using existing cable plants. IP-PBXs often work better by themselves on their own cabling infrastructure than those that are converged using just one cable to the desktop. I've claimed before--years before QoS, VLANs, network monitoring and assessments and other tools were typical--that by using available resources and installing the IP-PBX as a "phone system" is better for some customers.Voice uses very little bandwidth and this is a fact. Wiring does not need to be Category rated for voice to traverse copper. Now, if you want to run data over one copper cable with the phone and PC- of course you won't get ideal conditions and results vary unless that copper is Category -X but, I have attained and sustained 10 Mbps (half-duplex) voice/data (converged) connections in numerous offices where re-wiring isn't an option and the cabling doesn't carry a Category designation. These sites have okay connections, not great. Contrast this to Category 5, 5E, 6, 6A or 7 converged wiring and I'd say that your results vary too and to get good results of using one cable for the phone and PC connected to it, you will spend more money.
Large enterprises may want to seriously consider the idea of abandoning their structured wiring infrastructures because the mindset that it is easier to manage and costs less to maintain two infrastructures is not a fact. While these may be fighting words to some, for others depending upon how much of that Kool Aid they are on- it may seem unreasonable or impossible. Impractical? Hardly.
Recently, I meet with John Lambert of Valcom who is an RCDD (BICSI) and John's point to the campus folks we were presenting to was that "structured wiring is all wiring, and [that] doesn't mean it's all rated Category 3, 4, 5, 5E, 6, 6A or 7, and because it's not Category-X doesn't mean it's not an integral part of the structured wiring plant." Basically, the IT guys see one thing- data, and not that voice networks (cable plants) sustain many things: emergency notification, bells, clocks, timers, paging, talk back, relays, and more. These operate 24x7 with no help from IP and while many if not all of these functions can be modified or changed to operate over IP, what's the rush to do so with existing facilities that have been in place and working reliably for years? Be careful in responding, the wisdom of Peter Drucker may still prevail here.
Before getting to VoiceCon, I read over the news release from Phybridge. Of course I was excited since we've been installing IP-PBXs over existing "old" infrastructures for years. The Phybridge technology is very cool and I think it does offer an acceptable "bridge" between two worlds that are still very different.
First, you must understand the value of a cable plant. Unless you pull wire for a living- then you may not get it or have the same appreciation. That's okay, I don't get a lot of things until I experience them and then get an appreciation first hand. Nonetheless, wiring is an asset when done right- and this isn't an out to be used as an excuse by a guy saying the cable plant has too many issues. Inside wiring doesn't go bad. Wiring is bad because it's installed wrong, something was changed, neglected or damaged. On few occasions, lightning damage does bind the pairs (fusing) because of the incredible amount of heat produced.
Now, introduce a gateway between the voice wiring and the data world. The Phybridge Uniphyer gateway bridges these two worlds. The gateway uses one pair (one pair means two wires, two wires makes up a circuit) to connect any IP telephone to the voice cabling plant. That's right, one pair. Voice drops or cables that make up the cabling plant usually are 4-pair. This means four telephones from a TDM system can be connected to just one cable. The Phybridge box cross-connects to the MDF (Main Distribution Frame) or IDF (Intermediate Distribution Frame) or demarc (demarcation). The IP telephone on the user desktop plugs into a dongle (RJ45 - LAN connection). The other end of the dongle plugs into the voice faceplate (RJ-11 - one pair of wires). Follow the math and you will see that you can now connect four IP telephones to just one 4-pair cable.
The Phybridge gateway has two complementary LAN connections on each gateway, to connect to the data LAN. The gateway provides QoS on the voice network. The gateway also provides PoE to the IP phones and acts as a switch. The Phybridge solution acts as bridge between the existing voice plant (wiring) and the IP-PBX.
According to Will Harris at Phybridge, the IP phones operate at 1500 feet on the Phybridge gateway. The gateways are 48-ports and will scale in upcoming releases. There's a lot of interest from Canadian universities and while Phybridge is in start up, that's not something that I would count against them. Will has promised to get me a unit to test, and test I will.
What are the benefits?
First, the data network does not need a substantial overhaul. The traffic traversing from the voice cable plant (network) to the data (LAN) network will be management traffic, CTI (Computer telephony integration) traffic, and access traffic to the WAN. So the degrees of overhauling the data infrastructure are much less.
Second, why would you abandon your voice wiring? Unless it is truly a mess--and I will report more on this later--then you are throwing dollars down the drain by trying to maintain the mess than by abandoning it and moving on. Cable plants need assessing or valuating just like any other business decision.
Third, converged cabling challenges the company into a constant upgrade. The switches are now at 1 Gbps, but the phones are still 10/100Mbps. Then, the switches are upgraded to 10 Gbps; furthermore, what about PoE? There seems to be a constant bottleneck that goes unaddressed, or folks are wearing blinders. Why would anyone want to upgrade phones every 3-5 years or when the LAN switches need upgrading for more speed? Will the technology fix itself or improve, or will this awkward cycle continue as in the past?
Fourth- it's all about money. The Pybridge solution offers a lower cost to adopt and lower cost to own than converging the two networks (and we're only talking about copper wiring plants inside the building here). The "other networks" external to the LAN and phone system are what should be targeted for convergence. This is another point that has been overlooked. There are huge benefits in converging the "external networks (resources) not the cable plants unless you are going to build new. Building new, then, I'm pretty much onboard with convergence for Large enterprise but, no, not always or often enough for the SMB/E space, since there's still a void in supporting these smaller infrastructures and equipping them cost effectively with the right tools. (Packet Island may be changing this too)
Fifth- the voice traffic pretty much remains on the voice cable plant.
There are at least two people that are very excited about the Phybridge gateway- John Croce, CEO of Phybridge whom I spoke with at VoiceCon and me. Maybe I'm off my rocker and am missing something. I think at VoiceCon 2008, what was sorely missed by the masses above the noise about UC, was an incredible product that has great potential to bring more IPT systems into the fold.
