The Economic Benefits of Automating Capacity Optimization in IP Networks

Internet traffic continues to grow at a rapid rate year over year. Traffic growth is driven by recent technologies such as 5G and fiber to the home and also new applications such as AR/VR, cloud gaming, and video conferencing. IP aggregation and core networks carry the bulk of Internet traffic, and they are typically designed as mesh networks with multiple paths connecting an origin with a destination site. In traditional IP networks with shortest path routing protocols, networks are particularly good at rerouting traffic around link or node failures but are very ineffective at optimizing link utilization. Many communication service providers (CSPs) have used a brute force approach to network capacity planning by allocating extra bandwidth to links to ensure there is adequate bandwidth available to support unexpected bursts of traffic or short- term traffic growth. Typically, CSPs will engineer network links such that average utilization is 50% or less.

‍
As network links grow from 100GE to 400GE and larger it is becoming more important to use autonomous capacity optimization to optimize network link capacity. Links that are underutilized can support more traffic; links that are over utilization should support less traffic. In this paper we present a solution to this problem using the Juniper Networks Paragon Automation, which is a network automation suite that includes a Path Computation Engine (PCE) that simplifies traffic engineering, making it possible to leverage benefits provided by transport service paths, such as MPLS/RSVP, segment routing, and network slicing. It enables operation teams to manage strict transport service level agreements (SLAs) more efficiently and dynamically through automated planning, provisioning, proactive monitoring, and optimization of large traffic loads based on user-defined constraints. With this automation, operators can run their networks higher utilization while achieving predictability, resiliency, and SLA guarantees in service providers’, cloud providers’, and large enterprises’ networks. Our study shows that Paragon Automation can help operators increase average link utilization from 50% to 70% or higher.

‍
Autonomous capacity optimization is even more important today because silicon shortages have resulted in supply chain problems. Increasing network capacity requires CSPs to order new components that are delivered via the supply chain.

Delays in the supply chain could result in inadequate network capacity, causing serious network performance problems as well as SLA violations. In this paper we present the results of an ACG business model that compares two scenarios:

• With Paragon Automation

• Using brute-force capacity management

‍
The total cost of ownership (TCO) and return on investment (ROI) model compares the capital expense and operations expenses of a hypothetical network and shows significant savings using a PCE to optimize traffic engineering. The cost of network bandwidth is exceedingly high such that TCO savings in optimizing the network pay for Paragon Automation many times over. Our results show an overall TCO savings of 27%. We also show that even a minor increase in average network utilization of 0.5% will pay for the total cost of the investment in Paragon Automation.