Juniper Networks Saves Customers Millions of Dollars Using AWS Lambda

Founded in 1996 in Silicon Valley, Juniper develops AI-driven networking solutions used around the world. Its engineering teams must continuously innovate while meeting stringent constraints to deliver products and services that provide compelling value to its customers, including the majority of Fortune 500 companies and the world’s top 40 cloud and service providers. Because some companies use systems that date back to the early 2000s, new solutions must accommodate older tech while delivering modern functionality.

One of the challenges of running a critical, revenue-generating global network is resolving issues while minimizing downtime. Without device-level insight, engineers must assess massive networks that have tens of thousands of devices on them to pinpoint issues. “Previously, our engineers had to go back and forth with customers to identify problems, share network information, and develop solutions,” says Baum.

Juniper discovered new ways to deliver dramatic operational improvements through an ambitious project for a large service provider. A team of 15 engineers worked hand in hand with the customer to implement a deeply connected solution that provided insights that drove operational efficiencies to new levels. After 2 years of intense collaboration, both the customer and Juniper were excited by the results. The challenge was to cost-effectively scale the project’s capabilities across all customers while meeting a set of exceptionally complex requirements.

To unlock scale while providing comprehensive security for customers, Juniper chose AWS IoT Greengrass, an open-source edge runtime and cloud service, as a foundation for device software at the edge. And to deploy serverless architecture for its IoT solutions, Juniper pairs AWS IoT Greengrass with AWS Lambda, a serverless, event-driven compute service. By using AWS services, Juniper creates a secure connection from the cloud to on-premises components of its system without leaving residual data on the edge device. And because AWS Lambda works on a pay-as-you-use model, Juniper delivers a solution in which value scales with cost. “This is virtually the first time that a company can actually achieve the objective that we’ve all been working toward as an industry for over 20 years,” says Baum. Juniper is now delivering JSI to its global customers.

Over the first 12 months following Juniper’s launch of JSI, in late 2022, the company has focused on onboarding new customers. Juniper has seen that the solution is making a big impact. Customer feedback confirms that the solution is simple to deploy and use, as well as cost-efficient to maintain, which has been a major focus area for Juniper. The company aims to reduce mean time to resolution by at least 20 percent by empowering customers, partners, and Juniper teams with more information.

Using JSI, both Juniper and its customers get detailed analytics with which to proactively predict issues, plan device end-of-life operations, and keep critical infrastructure running. “A large customer who used JSI estimated they avoided 10 outages, which would have cost millions,” says Devin Petrino, product manager at Juniper. “We are saving time when our customers need our assistance, and customers are saving time and money on their planning.” In fact, customers’ costs have been flat even as they use the advanced capabilities of JSI.

By focusing on a pervasively serverless architecture and using AWS Lambda as a key element of the implementation, Juniper can securely automate the collection and reporting of customer device states, increase productivity across multiple organizations, and reduce the total cost of ownership. The serverless architecture has changed the way that Juniper engineers go about their work. “There’s something about serverless infrastructure that’s not readily apparent to people: it creates discipline in your development team,” says Baum. “It imposes design constraints and best practices that need to be adhered to if you want to get to the scale that we operate at.” By using AWS Lambda, Juniper has followed a highly consistent set of design patterns to minimize the tedium of constant code reviews or the nuisance of oversight due to unnecessary variance.