I remember in the 80s and 90s, when we Service Providers fought against the use of Packet Switched networks and IP.  To overcome the mandates of the data communications networks of our clients we created ATM (Asynchronous Transfer Mode) and said we now have a technology that meets the needs of data communications but also our need for control!  But we were wrong, ATM was just a costly (both in price and overhead) technology that came too late and was too restrictive.

With the advent of the Telecommunications Service Providers adoption of packet switched networks the mandate of using the Internet Protocol (IP) has become a necessary evil.  The driving factors for this has been the continued demand by our customers for higher bandwidth at a lesser cost as well as our (Service Providers) network usage demands with the advent of the mobile and broadband services.  Of course, prior to deregulation of the industry globally, we could charge a higher price and customers would either pay it or do without, which was the mind-set we Service Providers had!  But the market place has changed, deregulation has occurred, and customers are more demanding and we are now in a commodity industry!  Bandwidth is a commodity that can be shopped for!

With this revelation and the increase in broadband services, not only by our clients but also within our own internal networks, we had to adopt technologies and methods that allow us to reduce the cost per bit to obtain the greatest return on our bandwidth investment.

The continued use of circuit switching technologies, like ATM. PDH, SDH/SONet, has become a serious burden on Service Providers when considering the cost per bit aspect.  However, we (Service Providers) have come up with methods to allow Packet Switching over our circuit switched technologies through the use of VCAT, EoSDH, CENoSDH, etc.  However, this is only a band-aid, as we are not overcoming the technology constraints and overheads of circuit switched technologies, thus not achieving the goal of maximizing a reduced cost per bit.

But as in all things telecom – we are taking baby-steps to achieve the goal of a true packet switched network to address the predominant packet based traffic of today.  Numerous studies have shown that the traffic presented to us (Service Providers) by our clients and yes even our own network traffic, is packet-oriented, so the adoption of packet based networking technologies is obvious. But as previously stated, we, Service Providers, have always wanted to have a degree of control, not so much as to what is presented for transmission, but of the predictability of the transmission link.  The best way to achieve that is by having a fixed traffic path through our network regardless of traffic type or even traffic presentation.  It is easier to troubleshoot a predictable event then a sporadic even like packet based traffic.

With the adoption of packet based technologies into our network and recognition of Internet Protocol (IP) as the dominant traffic types, we (Service Providers) have had to accept the bursty (sporadic) nature of the traffic and the transmission link.  While accepting the bursty nature we still have this inherent need for the creation of ‘virtual’ circuits (paths) for traffic allowing us to maintain a degree of quality of service (QoS) that can be achieved easily with our circuit switched world.

Security has been a major concern of our clients for their networks and the associated traffic, thus they two preferred the lack of intelligence of the telecommunications networks previously provided based on ‘circuits.’   But with the increase is traffic volume and distribution of content, our clients need to have more bandwidth, more sites interconnected, but at a price point that is affordable for them.  Point to point circuits no longer adequately address our clients traffic needs.  The advent of packet based networks that provide multipoint level services with dynamic bandwidth is key but also still keeping it at a level that our (Service Providers) clients prefer to maintain traffic and content security.

The creation of Multi-Protocol Label Switching (MPLS) was a step toward the creation of a network that could provide the level of service to the clients while still allowing Service Providers with an ability to ensure traffic path predictability along with improvements on QoS features that previously packet switched networks with IP could not provide.  However, while MPLS supports layer 2 VPN (Psuedowire) it was designed to be optimal at layer 3.  But layer 3 provides a higher level of intelligence of traffic content to Service Providers that many clients do not desire.  Additionally, it potentially creates an environment where greater coordination of QoS features and addressing has to be coordinated between client and Service Provider.  As a result, Service Providers implemented MPLS however the dominant amount of the traffic being presented from clients and our (Service Provider) internal mobile networks is layer 2 traffic.  So in our core network we are able to process traffic more efficiently with the use of label switching we still have limitations on QoS and even a true multipoint structure as our label switched paths (LSPs) are unidirectional and did not support true muti-segment Pseudowire.

We recognized this limitation and the activities of creating GMPLS, T-MPLS, and/or MPLS-TP were started by various entities.  Ultimately MLS-TP, became the accepted standard of the three, thus the other two became somewhat incorporated into the final MPLS-TP standards.

MPLS-TP is a layer 2 standard that addresses the needs of security, switching, QoS, and multi-segment needs of our (Service Provider) clients while still providing us (Service Providers) with a predictable pathway while reducing processing time through our core network.  Additionally MPLS-TP allows for bidirectional Psuedowire within MPLS LSPs overcoming many of the traffic problems occurring with the IP/MPLS unidirectional LSP limitations.

The implementation of IP/MPLS within our (Service Provider) core network and MPLS-TP on the Access Edge network provides us with the greatest capability to properly support the packet based traffic that is being presented to us from our clients and our own internal network traffic (mobile & broadband) thus creating the proper platform for a packet based network.

Once we have implemented IP/MPLS and MPLS-TP structure the next big step will be the true obsolesce of our circuit switched transmission network (SDH/SONet/PDH) and implement a true packet switched network, like a native Carrier Ethernet Network (CEN) operating over Dense Wavelength Division Multiplexing (DWDM) optical network, a true packet based microwave network, a hybrid of DWDM/PON/Packet Based Microwave, and Passive Optical Network (PON) access network.

The time is now –  the technology is proven and stable.  Implementation of IP/MPLS, MPLS-TP, Carrier Ethernet over DWDM, PON, and or Packet Based Microwave is how we – Service Providers will be able to achieve the greatest cost per bit savings while still meeting the our clients and internal network demands.