Juniper Networks made a splash this week when it announced
its intent to acquire Wi-Fi vendor Mist Systems for $405 million, subject to adjustment and payable in a combination of cash and outstanding equity awards. The deal is expected to close in Juniper’s fiscal second quarter, ending June 30.
For those not familiar with Mist, the company was the first Wi-Fi vendor to use artificial intelligence (AI) as a way of automating the management of the network. The deal is highly complementary, as Juniper has a broad wired networking portfolio, but its Wi-Fi strategy had been a hodgepodge of partners including Aerohive, HPE (Aruba), Samsung, and Mist.
The purchase of Mist was something Juniper needed to compete effectively at the access edge. Juniper’s Wi-Fi partnership strategy might have enabled Juniper to check a box on an RFP that asked for wired and wireless solutions from the same vendor, but it would never have enabled it to provide a truly unified access edge.
More and more customers want an access edge where the wired and wireless network are managed as one. This has many operational benefits, as it allows network managers to do things like create a single set of access policies that can be enforced across the wired or wireless network. For example, it’s common for a guest to be blocked from the internal network via Wi-Fi but have full access when connected on a wired connection. With Mist, Juniper has the ability to create best-in-class experiences on wired and wireless, and across the WAN with its SD-WAN product.
Tackling Wi-Fi Problems
Unified wired and wireless is table stakes now, and Mist brings many advanced capabilities to Juniper. I think most people would agree that the Wi-Fi network has grown in value. What was once a network of convenience that offered best-effort services, is now the primary network for mobile devices and Internet of Things (IoT) endpoints that are often wireless only.
The problem is that Wi-Fi is notoriously flaky and unreliable. I’m sure everyone reading this has experienced a situation like being in an Enterprise Connect keynote and they’re able to Tweet, email, Facebook, and iMessage to their heart’s content. Then the first speaker comes out and everyone jumps on the Wi-Fi, starts uploading pictures of a nattily dressed Eric Krapf and everything stops working. I actually carry a 4G mobile hotpsot just for those occasions. Think about that -- I actually plan for the Wi-Fi to be problematic by carrying an alternate connectivity method. In 2019, that should not be the case, but it’s commonplace in hotels, schools, and other venues more often than not.
The challenge for network managers is that Wi-Fi problems are very difficult to troubleshoot. The example above is fairly simple to understand, as there are too many people connected; but with the right data, there are ways to fix that. It’s worse, though, when there’s only a couple of people on the network and things still aren’t working. Is it an issue with the client? DHCP problem? The application? RF problem? Some sort of configuration issue? The problem could be almost anything.
To highlight how big a problem this is, last year ZK Research ran a Wi-Fi troubleshooting survey and found that a whopping 60% of network engineers spent at least one day a week doing nothing but Wi-Fi troubleshooting. Even more alarming is that the primary troubleshooting tool is packet capture, which any network engineer will tell you is the tool of last resort. But there’s a lack of good Wi-Fi troubleshooting tools, so here we are with packet capture.
This is where Mist excels. Instead of building a tool to problem solve Wi-Fi, Mist uses AI to create a network that can run itself and make Wi-Fi more predictable and reliable. For example, if there’s a channel conflict, the Mist system will see that and fix itself.
Also, about a year ago, Mist debuted
its Marvis virtual network assistant (VNA) for operations and integrated helpdesk, which is powered by its AI capabilities that include natural language processing (NLP). Marvis is a play on words where “Mist” is combined with “Jarvis” of Iron Man fame. Using Marvis, network managers can be like Tony Stark and ask Marvis things like, “Why can’t Beth Schultz’s iPhone connect to the network?” or “Tell me the five locations with the poorest Wi-Fi performance,” or “How many clients are connected to the guest network?” Marvis is like having a virtual Wi-Fi expert always at the beck and call of network operations, able to quickly connect the dots between mobile devices, wireless networks, the wired network, and IoT domains.
This Mist solution is delivered from its modernized, micro services-based cloud, so it’s well set up for the future. I expect the Mist cloud will eventually be the primary management platform for all of Juniper’s enterprise products. Within the next five years, I fully expect AI-based solutions to supplant humans when it comes to troubleshooting, tweaking, and tuning of the network at every point.
The $405 million purchase price seemed low to me. No revenue number was provided for Mist in Juniper’s press release, but my Valley sources tell me it’s in the $30 million range, with Mist expecting to double that this year. At $60 million, the $405 million represents a multiple of 6.75x revenue, which seems modest. Also, Mist has some outstanding lighthouse customers such as Amazon, The Gap, Disney, and Verizon Enterprise, which are often just as important as revenue.
With Arista having bought
Mojo Networks, and Extreme Networks having rolled up
Avaya’s networking business, Brocade data center products, and Zebra Technologies’ Wi-Fi business, the network edge is significantly more competitive than it was just a few years ago.
The network edge is where the action is, as it’s the place where people work, shop, and learn, and it’s where IoT devices connect. Juniper had tried to go to market via partnership but that has many challenges. The acquisition of Mist gives Juniper a best-in-class, AI-driven, cloud-based Wi-Fi solution that should become the cornerstone of its overall enterprise network strategy.