It is a lot easier to provision bandwidth in a Software Defined Network (SDN) architecture than in the old TDM-based world. In the past, increasing bandwidth was mostly by 1.5Mbps increments and it would take at least 45 days to 90 days to process an order and get the service up and running. Today, Ethernet access provides a much better design and flexibility and if coupled with the right Network Controllers, can provide you great relief as you are able to manage your network dynamically.
Because new services providers’ architectures are finally moving to Ethernet switched access network technologies, these architectures are becoming a lot more economical too. Let me walk you through a customer use case example where a $7.7M investment to upgrade the network to support this new architecture generated $25.8M hard-dollars savings in a 5 year period. So, the proposed network upgrade was self-funded. The result was a simpler and more flexible network architecture, they improved levels of service, maximizes network throughput all the while reducing their overall network costs going forward.
Current Network State
This is an enterprise with 356 locations across California. They have a large VPN MPLS network connecting their office locations sites to three data centers. Here is an overview of their initial network state and key metrics for this network:
In summary, they have 96% of their VPN network still on TDM technologies ranging from OC-3s and NxT1s based access. The rest supported by MetroE or Ethernet-based access. As you can notice, there is also a separate voice network, which is 100% TDM-based. Below you can see a small dashboard visualizing two key areas used to drive the new network design.
Each organization uses secure private access in its core network giving access to common and individual enterprise applications available at the data center locations. With their current method of operations, each office location required separate IP networks, i.e. VPNs ingress were set up by departments in each location, resulting in many network management and administrative challenges (Departments x Networks x Locations). So the main premise for the future solution design was to eliminate these operational constraints, minimize the risk associated to this inflexible TDM-based network architecture, and ideally reduce the overall cost of the network at the same time.
Future Network State
Based on the requirements above, we proposed an SDN-based architecture which eliminated physical restraints, so different department who were in same or separate offices and branches could have their own private clouds, simplifying network provisioning and administration through a centralized network controller. Which also gave the customer dynamic flexibility on how they run their bandwidth in a way that was not possible before.
This design leveraged Ethernet switched-base network access technologies from the service providers, so locations who were part of the same local area and transport areas (IntraLATA) can create virtual private lines (VPL) dynamically, including capacities from 100Mbps to 10Gbps within those branches clouds (see pic above).
At the core, a 10G network core with QoS interconnecting different local access and transport areas (InterLATA) posed a better and more reliable and redundant configuration to interconnect clouds at different group levels and within network clouds from each of the data centers.
While aligning this architecture we modeled several financial scenarios and fined tuned the design resulting in this final awesome business case.
As you can see in the above graph the $7.7M investment necessary to implement the future network architecture (including hardware, software, and services) generates $25.8M hard-dollars savings in a 5 year period. This resulting in $18.1M of net savings!
So, this type of architecture provided network access bandwidth which can now easily scale as needed (100Mbps to 10Gbps) with better quality and considerable improvement in throughput and adaptability to future demands. Moreover, this single IP network clouds design assure all type of bytes (voice/data/video) to be managed simultaneously with less complexity, no points of failure, with fewer risks, and a more secure way for sharing information within internal and external organizations (VPLs). Also, a significant uplift in network capacity from 17.4Gbps to 230.9Gbps (13.3x) was achieved. And all of this self-funded! Same budget (or less) and a lot better network.