Packet Transport – It's Time To Pay the Bills
Posted by Gil Epshtein on 5 Dec 2012
I think that each and every one of us has experienced at least once the “cash register crisis.” I am referring to that moment when you get to the cash register to pay for all the “must have” items you have just picked up, but then find out that the total bill is much higher than you anticipated.
Looking at the packet transport market, I have a sense of déja vu. As some veterans still recall, packet transport was introduced to lower the cost of transported bit so operators could support the huge increase in packet traffic in a cost-effective manner. In reality, though, once packet-based transport equipment was equipped with all the relevant transport grade features, it became much more expensive than the equivalent equipment used for enterprise applications. But not only direct capex was off target. Operators soon realized that running such packet-based network, or more specifically an IP-based network, is a complex and thus also costly experience. Last, it became clear that, despite the explosion in data traffic, TDM-based services are here to stay, and any transport solution will have to support them as well. Supporting these TDM services over packet infrastructure through circuit emulation (CES) is also a costly experience.
Bottom line: building a carrier-grade packet transport network is much more costly than anticipated, jeopardizing the original goal of lowering cost per bit. As a result, operators are re-evaluating their requirements and assumptions. One of the basic assumptions being evaluated is whether packet transport requires IP and L3-based equipment. Operators have found out that, by filtering the requirements from the supporting equipment, the same result can be achieved with a much cheaper MPLS-TP- based solution and without compromising on any of the requirements. Moreover, as the MPLS-TP-based solution can be managed by a centralized and easy to use GUI-based NMS rather than rely on a distributed and complex control plane, not only capex is lower but also opex is reduced substantially.
Another element being evaluated is the way to handle the remaining TDM services. The common belief was that the most efficient way is to move all traffic to a single packet based converged network. Eventually, we realized that the key for efficiency is to use a single network element (NE) that can handle both TDM and packet traffics in a flexible manner, sometimes natively and sometimes using encapsulation (CES or EoS) and with the ability to change over time according to the nature of your traffic going forward.
So just as with our “cash register crisis” example above, the point is not to “put things back,” but , really re-evaluate your “must have” (requirements), detach the list from any pre-conceived idea you might have regarding a specific item that can address your requirements, and open your mind to new point of views.
That was our thinking when we came up with the NPT (Native Packet Transport) approach - simple and efficient carrier-grade packet-transport solution for low TCO, fulfilling the original promise of packet-based transport networks: lowering the cost of transported bit.
So take your basket, enjoy the goods and avoid the panic when you get to the cash register.
Topics: Packet Networking