As service providers look to grow subscribers and revenue, the virtualization of network functions is key to reducing costs and enabling new service development. Currently, customer demand for bandwidth is growing, and operators must provide additional network capacity while investing in new services to keep pace with subscriber expectations and drive profits.
As SDN/NFV adoption approaches critical mass, these technologies propel the impressive growth and scale of large content and OTT providers, with market estimates for 2022 ranging from $54 billion to $168 billion. The CORD initiative (central office re-architected as a data center) offers a framework for SDN/NFV implementation and provides a number of benefits, such as helping operators to make networks more agile and cost-effective, as outlined previously. In this post, we discuss key considerations for CORD implementation.
In the ebook this is ‘Central office with data center pod (green)’
CORD doesn’t only require different hardware and software to traditional solutions but also needs an architectural change. Architectural evolution is necessary to enable new ways of managing and delivering services, and when implementing a CORD infrastructure, there are some essential considerations for SDN/NFV success:
Retooling facilities to accommodate these new technologies is crucial. The physical CORD adoption in a central office requires a data center-like architecture which is often deployed in a data center “pod”. This is a separate space in the facility where processing takes place. These pods are equipped with a high-density leaf-spine network, optimized for cloud-oriented east-west data flow, deployed on the network edge. The pod architecture also takes advantage of cost savings from lower-cost multimode fiber optics.
The cabling infrastructure inside the pod must be carefully planned by engineering teams to support multiple migrations, in an environment where migration to higher speeds is common. Indeed, lane capacities are expected to keep increasing, reaching 400G by 2020 and enabling the next generation of high-speed links for fabric switches. There are several factors driving the surge in data center throughput speeds, including a 20% a year increase in server densities – and processor capabilities are increasing at a similar rate. It’s also important to take into account the industry’s Ethernet roadmap for multimode optics. Multimode transceivers provide a cost saving when compared to singlemode and are often the first choice for data centers.
High port density is essential. The mesh of fiber links in a leaf-and-spine topology creates a high-capacity network resource (a “fabric”) that is shared with all attached devices. All fabric connections run at the same speed so the higher the speed, the greater the capacity of the fabric. Fabric networks require a large number of connections, especially in the switch layer. Properly designed fiber apparatus should provide higher connection density while avoiding long-term management issues so that fast, easy deployment can serve large-scale environments where moves, additions, and changes are required.
SDN/NFV are increasingly key to new network architecture, leading to greater efficiencies in terms of management, operations and cost. Engineers that introduce infrastructure “building blocks” enabling easy and fast migration to higher speeds will ensure data center pods are able to cope with the exponentially increasing demand for higher capacity and ever-changing network requirements.
Ultimately, the CORD initiative enables demand-based, value-added (cloud) services which means that operators can quickly and effectively introduce new services and innovate their business models.
You can learn more about considerations for the modern data center here.