Automation is one of the big rising stars of networking. It has attracted huge industry interest of late, with Light Reading dedicating a whole category to the topic and many industry events, including Layer123’s SDN NFV World Congress, hosting significant automation tracks. Carriers are building requirements around it in their RFP processes, and analyst firms have entire practices dedicated to the topic.
As a business founded on the basis of enabling IP/Optical network automation, we’re very enthusiastic to see the vision of automation starting to gain traction with global service providers, media, and analysts alike.
Automated service-aware networks (the goal of most major carriers), allow network resources to be pooled and allocated dynamically, based on the specific needs of each service. This results in improved utilization/efficiency, while supporting assured service protection and performance (e.g. latency), and enabling new opportunities to grow revenue through service differentiation. This represents a fundamental departure from today’s networks which are, at best, aware of aggregate traffic requirements.
A successful automation strategy however first requires an accurate real-time understanding of the network. Without this, network controllers may initiate changes that lead to errors that result in network outages and service failures.
In this post, we focus on Network Intelligence – the essential first step towards network automation. It provides a 100% accurate view of the converged multivendor, multi-domain IP/optical network that is an absolutely fundamental requirement for any automation plans.
As we shall see, Network Intelligence also enables many powerful use-cases – delivering both cost savings and revenue growth benefits well in advance of full-automation.
Network Intelligence – Discovery & Analytics
Network Intelligence closes the loop between planning, operations, and sales groups such that each team has the same 100% accurate view of the network.
During the network discovery process, a central IP/optical controller (such as Sedona’s NetFusion™) extracts data from the network, and then abstracts and visualizes a converged real-time view of the complete multivendor, multi-domain IP/Optical network with all its relationships and dependencies. This network-wide view of topology, traffic, and services delivers unprecedented visibility, and forms a “golden source” of accurate real-time network data that other applications and OSS/BSS systems can make use of.
We’ve found that carriers are typically amazed when they see the results of network discovery – it’s the first time they have ever seen a complete and accurate view of their network, and represents something of a dream come true for a network planner or operator.
With the accurate network data in hand, NetFusion then applies built-in data analytics to identify anomalies in the structure and behavior of the network.
One of the most profound aspects of the discovery and analytics process is that it doesn’t require SDN controllers or latest generation IP or optical systems to work. The Network Controller can extract the data it needs not only from SDN controllers, but also from legacy EMS/NMS systems, and, if necessary, directly from IP or optical network elements. This allows carriers to start to enjoy the benefits of converged network control with existing equipment, and becomes a bridge to the more holistic benefits of SDN over time.
Now that we have all this accurate intelligence from the network, let’s look at what can we do with it. Here are just a few high-value use cases:
Operational Efficiency Improvement Use Cases
Assured Network Inventory (and Reduced Order Fallout): Having accurate, up-to-date network inventory data is critical to delivering new services. Even minor discrepancies can result in order fallout that is both costly and impacts customer satisfaction. Inventory records are typically less than 100% accurate however because multiple activities (operations and maintenance) are constantly touching the network. When mergers and acquisitions happen, more inventory shifts occur, likely across vendors and optical domains.
With access to accurate real-time network intelligence however, inventory systems can be checked and validated against a known accurate source, with errors easily corrected in seconds or minutes. Regulatory audits, once a laborious, time consuming manual process, can now be completed quickly and accurately.
Fact-Based Planning: Network planners monitor network topology and traffic in order to make critical assessments of how many IP and optical links to provide and where best to place them. Without current and accurate multilayer knowledge, this process is complex, requiring data collection from multiple platforms and intense collaboration between multiple teams.
With a pool of real-time network intelligence, planning tools can make use of 100% accurate data extracted directly from the network itself. This takes care of the heavy-lifting and allows planners to focus on fast, accurate decision-making.
Optimized Maintenance: For most service providers, the IP and optical operations are often not coordinated. As a result, maintenance activities in one layer often happen without the operations team of the other layer being aware. An optical maintenance team for example, may not be aware of the IP links that traverse a given optical link, and could create an unanticipated service outage when the optical link is shut down for maintenance.
With access to real-time network intelligence, operations teams can see the full impact of taking links out of service because they can easily see the dependencies between the network layers. As a result, maintenance activities can be planned with confidence, with traffic re-routed in advance of the maintenance window so that there are no service outages.
Assured Diversity: High-availability services typically require a pair of MPLS LSPs that are routed over diverse fiber links. If one of the LSPs fails, its traffic switches quickly to the other one so high-availability SLAs can be maintained.
But what if the two IP links unknowingly share a segment in their optical paths? This risk is very real because many IP links are physically routed through a smaller set of fibers and because cross-layer awareness is very limited in existing network topologies. A single fiber cut might expose the fact that a pair of connections is not actually independent, resulting in SLA violations.
Real-time IP/optical network data provides visibility to exactly how each IP link is mapped to an optical link. Diversity problems can thus be flagged, and when appropriately configured, IP controllers may be notified directly, allowing them to take appropriate corrective actions.
Revenue Growth Use Cases
In the previous examples, we’ve focused on cost reduction and operational efficiency improvements, but network intelligence also provides some significant opportunities to grow revenue. This is particularly important in today’s highly competitive marketplace.
Shorten Time-To-Revenue: Earlier we described that accurate network inventory data results in reduced order fallout. A direct consequence is that the time interval from order placement to implementation and starting to recognize revenue is shortened. In many cases this may result in a month or more of additional service revenue.
Having an accurate real-time view of the network also means that sales teams can improve targeting and focus their attention on opportunities and areas where capacity and/or services are known to be available. The result again is shorter time-to-revenue.
Increase Sales Close Rates and Revenue-Per-Service: Accurate network intelligence means that the carrier sales team can have access to accurate Design Record Layout (DLR) during the presale and post-sale process to prove diversity, latency, and optimal path. As a result, they can increase their competitive advantage and sales close rates by offering performance guarantees (typically latency and route diversity assurance).
With verifiable performance guarantees, carriers can also charge higher rates for premium services that they were previously unable to offer.
Implementing an effective network intelligence solution is a powerful first step for carriers in their journey towards fully-automated service-aware networks. It not-only delivers high value today with existing legacy systems, but also forms a valuable bridge to controller-based networks. Finally, it allows carriers to develop a degree of familiarity and comfort with allowing their networks to be under software control.
At Sedona, we’re deploying our NetFusion IP/Optical Controller in Tier 1 service provider networks today, specifically for its capability to discover, abstract, and visualize the converged IP and optical network view and to provide the golden source of network intelligence we’ve described in this post. It’s now a proven model and we’re excited to be a part of our customers’ evolution to more fully-automated service-aware networks.
Note: In future posts, we’ll delve more deeply into the business case justifications for many of the use cases outlined above.