Most hyped network technologies and how to deal with them

The appeal of promising network technologies can be jaded by pressure to adopt untested ideas. When I look over the comments I’ve gotten from enterprise technologists this year, one thing that stands out is that almost three-quarters of them said that entrenched views held by company executives is a “significant problem” for them in sustaining their network and IT operations.

“Every story that comes out gets me a meeting in the board room to debunk a silly idea,” one CIO said. I’ve seen that problem in my own career and so I sympathize, but is there anything that tech experts can do about it? How do you debunk the “big hype” of the moment?

For starters, don’t be too dismissive. Technologists agree that a dismissive response to hype cited by senior management is always a bad idea. In fact, the opening comment that most technologists suggested is “I agree there’s real potential there, but I think there are some near-term issues that need to be resolved before we could commit to it.” The second-most-cited opening is “I’ve already launched a study of that, and I’ll report back to you when it’s complete.” There’s usually a grain (yeah, often a small grain) of truth underneath the hype pile, and the best approach is to acknowledge it somehow and play for time. Hype waves are like the tides; they come in and they go out, and many times management will move on.

Now let’s consider what technologists say are the specific big hypes that relate to networking, and what they recommend to get past them.

1. AI-driven network automation

The number one hype on the list these days is surely AI automation of network operations. Just over three-quarters of CIOs said that they were asked about “taking advantage” of things like ChatGPT to “run” their networks. They agree that the best approach to dampening pressure is to advocate AI incrementalism.

I saw one presentation a CIO gave to senior management, and it had four main points. First, AI is evolving rapidly and even now shows great promise, but the fact that AI makes mistakes today is well documented. Second, if AI were given control of the network, and if it made a mistake, it would be difficult for the network staff to determine what it had done and why it went wrong. Third, the best way to deal with the first two points is to introduce AI in an advisory role, in a confined area of the network.

Lastly, a project should be undertaken to identify an area of their network where there is a single dominant equipment vendor that has an integrated AI advisory tool, and integrate that tool into their network operations practices. The project should 1) assess the value of the tool by comparing operations response times to issues with the same network segment but without AI, 2) develop recommended procedures for optimum AI use, and 3) monitor AI evolution to assess when AI might be able to actually be trusted to take actions. At the minimum, this step should be taken only if the steps AI takes automatically are journaled so recovery from a mistake by AI is facilitated.

2. Private 5G

The technologists say the number two hype they have to deal with is private 5G. A bit over two-thirds of technologists said that they were asked whether the company should be using private 5G, and almost half said someone had suggested it should be considered in place of WiFi. Technical specialists said the best approach to this “suggestion” is to point out that WiFi supports the great majority of devices the company already uses, and that 5G should be considered a parallel connection option where WiFi doesn’t reliably reach. This is most likely to be the case for specific verticals, including public utilities, transportation, and government, and, in particular, in new IoT applications rather than applications that require connecting existing devices to a network.

The project suggestion here is a facility audit, which would look at each company location to determine if it is properly served by WiFi and if there are proposed applications that would justify adding private 5G as a connection option. For those sites where a possible private 5G mission is identified, the project would then assess the options for creating a private 5G network, including the use of cloud-provider private 5G services or network vendors with extensive private 5G experience, preferably based on interviews with other similar companies that have considered/adopted private 5G.

3. Open networking

Number three on the hype list is “open” networking. One network manager told me that his CFO had read that adopting white-box devices and open-source software could “cut our capex in half next year.” Almost half of all CIOs said that they had to justify not converting to open technology in response to an inquiry. Given that open networking is widely accepted as the lowest-cost option, it’s difficult to convince management there are issues in realizing the benefits. Those issues relate to the challenge of evolving from a current proprietary network infrastructure to an open infrastructure.

The project suggestion here is to start with an audit of the expected useful life of specific segments of the network, focusing on locations where switches are used (like the data center) or where a technology shift is being considered (the use of SD-WAN to expand the scope of the company VPN). The best place to introduce open-network technology is a place where all equipment in a part of the network can be changed out at one time. Otherwise, new open-model devices will introduce changes to operations practices and the benefits of a few open devices will be unlikely to compensate for the impact of these changes.

4. Zero trust network access

Hype four is zero trust. Security is justifiably hot, and at the same time there never seems to be an end to the new notions that come along. Zero trust, according to technologists, is problematic not only because senior management tends to jump on it without thinking, but because there isn’t even a consistent view of the technology being presented. “Trust-washing,” said one professional, “has taken over my security meetings,” meaning that they’re spending too much time addressing all the claims vendors are making.

Technologists say the best approach to a project to address this hype issue starts by redefining “zero” trust as “explicit trust” and making it clear that this means that it will be necessary to add tools and processes to validate users, resources, and their relationships. This will mean impacting the line organizations whose users and applications are being protected, in that they will have to define and take the necessary steps to establish trust. Zero-trust enhancements are best implemented through a vendor already established in the security or network connectivity space, so start by reviewing each of the tools available from these incumbent vendors. Get representatives of line departments involved! A zero-trust investment is wasted if validation of users and resources isn’t established and maintained, and in fact a failure in this process can actually lull users into believing they have better security than they do.

You can see that the recommendations of tech professionals for any and all management-driven hypes is not to try to say “It’s all hype” but rather to launch a project that’s designed to show the real value and real risk of any suggestions aimed at adopting a technology that’s been over-promoted. Remember that under every hype wave there’s a grain of truth sand, and uncovering it can benefit both your company and your own career.

Source