The key takeaway from NXTWORK 2019, the annual Juniper Networks event for network ‘gamechangers’ in Las Vegas, is that all the pieces of Juniper’s acquisitions are coming together. The dots are connected. Acquired companies such as Appformix and Mist, which were seemingly standalone solutions, are now rapidly evolving into becoming a fully integrated offering.
Juniper convinced me during the event that this will significantly benefit the simplicity of managing your Juniper network, your security and even your multi-cloud environments.
Obviously, Juniper Networks will continue to use open standards and embrace interoperability, aiming for an end-to-end proposition that fully supports the enterprise.
Here’s what you missed during NXTWORK 2019:
1. A new era: The Artificial Intelligence (A.I.) driven enterprise
By integrating AI into enterprise network and security, Juniper Networks announced a new era. AI-driven networks enable operations to automatically detect and resolve issues, even before they turn into incidents. As Juniper receives metadata from all connected customers, the machine learning engine will become exponentially smarter and faster at resolving problems than a human could ever be. Juniper is aiming for a self-healing and self-driving network, which enables network and security engineers to act more like architects. Because when AI drives the enterprise network, many daily/ routine actions will decrease significantly. Jobs that currently focus on “keeping the lights on”, could eventually be replaced by A.I. to ensure stability, reliability and uniformity of operations.
2. Multi-cloud operations
It has become more and more challenging to manage on-premise network infrastructure, as well as multiple (public) cloud IaaS (Infrastructure as a Service) platforms. The reason for this is that every platform uses its own interface and coding to deploy and operate. This is where Juniper’s Contrail Enterprise Multicloud comes in, fusing all platforms together. It acts as a single pane of glass to operate multiple platforms, ensuring consistency in deployments and operations. Juniper’s mindset has always been to create solutions based on open standards, therefore interoperability is ensured by using all available APIs and protocols.
"Jobs that currently focus on “keeping the lights on”, could eventually be replaced by A.I. to ensure stability, reliability and uniformity of operations." - Jan Sixma, Sales Executive at Infradata
3. Get a real time view of what’s happening on the network
The acquisition of Appformix gave Juniper the possibility to trace applications sitting on the network in real-time. This capability is now fully integrated into Contrail Service Orchestration and is rebranded to Contrail Insights. Troubleshooting and rerouting applications reduce complexity significantly, because all network information is instantly available. From an application point of view, this could be bottlenecks, connection failures, congestion, package loss, etc. With Contrail Insights the thresholds for an application to run on an environment can be monitored and the operations team will be able to instantly see possible problems with applications that negatively impact the user experience, even before end users have the ability to raise a support ticket.
4. The Marvis A.I. engine will be unleashed onto the wired network
Juniper’s recent acquisition of Mist Systems was not purely about completing the network portfolio. What became clear during NXTWORK 2019 is that one of the reasons for Juniper Networks to buy Mist was Marvis: the A.I. engine Mist is using to ensure a trouble-free and optimal WiFi network performance. Most wireless networking issues will have already been seen and solved by other Mist users, allowing new customers to benefit from this experience. Previously, every single customer had to endure the long and tiring troubleshooting journey, moving from a symptom to the origin of the incident and eventually solving it by trial and error.
Marvis fully automates this process. When using Mist for WiFi, Marvis already solves more than 50 percent of errors without human/manual intervention. Juniper believes that this percentage will eventually get close to 100%. With this “A.I. problem solver” now also available for the wired network, it could also be possible to perform daily tasks, which saves manpower and introduces consistency by acting as a target blueprint for the network. By creating a template, the configuration of the components can be "pushed" to guarantee consistency.
5. A threat aware network with SecIntel (Security Intelligence)
Since malicious attacks do not only originate from outside the organisation, it makes sense to upgrade defences within the network itself. By enabling SecIntel on all Juniper product lines, from the Mist Access Points to EX & QFX switches, MX routers and SRX firewalls, Juniper improves their security significantly. Using its Policy Enforcer enables immediate automated action when needed, which is a lot faster compared to required manual operations from the SOC. The automated solution provides improved defence systems against the possible spread of malware or the impact of a hack.
6. Optimising for Enterprise and Cloud
It’s clear to me that recent developments at Innovation Way 1133 in Sunnyvale (the headquarters of Juniper Networks) are strongly focused on Enterprise and Cloud environments. Although the announced increase of 400Gbps in throughput on some Juniper platforms is not really relevant for most enterprise customers, most of the fancy shimmer and shine at Juniper is not focused on ISPs.
My impression is that the next phase for Juniper will be to gain market share within the enterprise and cloud segments, competing against Cisco and HPE Aruba by providing fully integrated and open stack solutions that can be easily adopted by customers, step by step, to avoiding sudden big bang migrations.
After seeing demos of several solutions at NXTWORK 2019, I am convinced that customers will transform their current infrastructures into future proof self-driving networks at their own pace, with the reassuring thought that they can rely on their A.I. engine. That’s why I believe we will increasingly be looking at AI-driven enterprise networks in the years to come.