IP networking hasn’t changed since the mid ’90s. For the first time in 25years, as software-defined networking (SDN) takes off, disruption is occurring in the networking field. SD-WANs are a start, but they’re more of an evolution of IPSec tunneling versus a full disruption.
Unlike chemistry or the hard sciences where there are fundamental rules, the architecture guidelines on how to design, build, and operate a network changes every 10 years or so. Below are eight disruptive changes in how networks will be built in the future.
- Private 5G as a Wi-Fi X replacement — For new construction of a large building or campus, 5G is faster, better, cheaper, and more secure than Wi-Fi 6. For companies that run Power over Ethernet and 100/1000Mbps to their access points, 5G vs. Wi-Fi 6 will need a careful analysis since Wi-Fi 6, and 7, are best served by 10G fiber. For companies that just need more wireless capacity and have existing Wi-Fi infrastructure, Wi-Fi 6 makes sense.
- Prediction: 5G will move into the WLAN market.
- Ability to use multiple networks concurrently — From a mobile device, users can tap into multiple 4/5G and Wi-Fi networks concurrently instead of having to roam from one to another. Electronic SIM, or eSIM, cards allow use of services from multiple cellular providers. For power mobile users, the ability to use multiple networks concurrently to increase bandwidth, lower costs, and provide another level of security is worth it.
- Prediction: Faster, better, cheaper, and more secure wins in the marketplace, and power consumers will demand the real-time freedom of using whichever service they want, when they want.
- Zero trust networking — The zero trust security model is gaining momentum, as evidenced in the term’s wide use at this year’s RSA Conference. With zero trust networking (ZTN), no packet gets onto a network without prior authentication and authorization. Thus, all malware and malicious traffic is stopped at the edge of the network versus in the middle, or on your virtual doorstep.
- Prediction: ZTN is the only way to truly secure networks and networking security will become the top priority of next-generation networks over and above being faster, better, or cheaper.
- Doing away with IP-based routing — IP addresses themselves aren’t going away anytime soon, but the importance of them will fade as SDN takes off. With SDN, the control and management planes will use words rather than IP addresses to define routing and security policies and integrate with DevOps tools and applications.
- Predictions: 1) IPv6 will continue to be of little value to any organization, and 2) future routing and security protocols will be based on words as software eats the world.
- Open source — Open source has disrupted every area of IT, with networking being no exception. As I wrote in a July post, networking has become ripe for open source adoption. The big cloud companies all built their own networking solutions using commodity hardware, open source, and some of their own code, because nothing on the market could meet their needs cost effectively. Plus, in a software world, routers can scale horizontally, so gone are the days of paying a million dollars for a router that handles 1Tbps of throughput.
- Prediction: Open source is coming in networking. Embrace it and learn how to code. The CLI is dead.
- Bringing the cloud to the enterprise edge — Networking and security are the top barriers to cloud adoption. In order for a cloud company to deliver end-to-end quality and security effectively, it must manage the network between its applications and users. Microsoft, for one, has embraced this strategy, as witnessed in the Azure Virtual WAN service. Equinix, with virtual colocation, is driving networking closer to the edge of companies versus making companies backhaul their traffic to just a few colocation sites. Many of the UCaaS providers have adopted SD-WAN to help ensure quality.
- Prediction: Networking is the bane of any cloud company, but more and more will get into networking or partner for it to deliver their service directly into the enterprise.
- Endpoint multicast — The killer application for the Internet is “look at me now” — Facebook, Instagram, YouTube… these are all built around the idea of monetizing content users want to disseminate, often while trading off their privacy. True freedom will come when users are able to send content directly to their followers. This will require multicast directly from the endpoint.
- Prediction: A distributed, more private user-to-user service will replace Facebook.
- 5ms network latency for apps — Virtual/augmented reality, auto-driving, and next-generation interactive applications will require that users be within 5ms of the applications they’re consuming. This means no more VPNs, and no more network backhauling to a data center.
- Prediction: The trend of distributing Internet exchange points will continue — eventually there’ll be one in your neighborhood.
One thing that isn’t on this list is intent-based networking. While automation works on fixing one broken component or a single specific change, networking as a whole is a unique system for every company. Fully understanding how one change can impact the entire system requires human knowledge.
Baseball legend Hank Aaron became the world’s leading homerun hitter by going to bat often…. That’s to say, a lot of my five-year predictions come true, not because I’m all-knowing, but because I make a lot of predictions and a few come true. So, if your company invests big in WiFi 6 or IPv6, or signs a five-year contract with Vendor XYZ, consider yourself forewarned. Change is in the air, and not all long-term assumptions will clear the fence.