Currently, the networking billing models for many technologies in the networking sector bill on a single instance license basis, that is, you pay per each site/device/user/instance that you have. For example, this means a separate SD-WAN license for each site or device, with a new cost incurred every time you add a new location. These licenses are often tied to bandwidth use as well, with each site given its specific bandwidth limits and surge capabilities.
This strategy made sense for a long time as most network analytics are retrospective. You knew you had a bandwidth problem when devices did not work. You pulled network analytics after-the-fact to make determinations about your network. However, with the advancements in network technology, more reliance on cloud services, analytics have become a real-time function of newer network architectures.
Solutions already offer real-time reporting on bandwidth usage on a per-application basis. This approach allows these solutions to make intelligent routing decisions for each app and to report on individual performance. As you can imagine the usage profile of each app changes throughout the day and week, which means that an app could have drastically different use from when it checks for updates on a Monday morning, compared to its use for backing up the database on a Friday evening.
Under the old model of billing, the backup usage would be considered the peak that you would have to pay for, for every site using that application. Yet is it fair to ask a company to pay for 500mbps for an entire week when they only need it for 2 hours a week?
This conversation, we predict happening more and more as enterprises gain a deeper understanding of what their network needs are. Also matched by development in automation and network intelligence, which could if it wanted to interact with billing models to scale them up or down based on needs automatically.
Why is this needed? Well, paying per site or license becomes a drastically inflated number the larger your site/user base is. This year’s pandemic has changed home working into something that roughly 1/3 of employees across developed markets will be doing on an ongoing basis—in some cases, changing the number of “business sites” from 20 to 2000. How to treat these new users who have come out from behind the site device/license and now are individual sites on the network is a topic for large amounts of discussion. Some of these individuals may, at specific times, require the same amount of bandwidth as a small enterprise site when dealing with large data-led projects, but probably not on an ongoing basis. Should a company then pay for the high flex limit for all these employees? Or should there be a system in place for repurposing bandwidth/devices/licenses to provide for specific needs at specific times, at a flexible price?
I believe we are coming to a change point in networking billing. Notably, a few vendors have already started talking about this. I heard 128 Technology (now officially acquired by Juniper Networks) talking about its plans for a flexible-edge network deployment which included more flexible billing at the recent SD-WAN and SASE Summit run by Upperside Conferences. Similarly, I have heard from other vendors who are planning to include billing and network upward flex limits by region rather than by device—allowing any site within that region to flex to the regional limit, rather than setting these per site.
Other vendors I have spoken to have been discussing, rather interestingly, moving over to a pay-per-usage model, aligning these vendors with the large cloud providers. They charge per MB of data that transits their networks. As these vendors are often hosting their solutions on these platforms, using them for transit as well makes sense, and creating a model that provides a single cost per MB including both transit and the cloud networking solution.
One question that keeps coming across my mind and often comes up in discussions with the team here at Cavell is what happens when you agree on a shiny new contract with a set bandwidth amount. You deploy all these new intelligent solutions, and it tracks the network for a month, and you realise you are using much less bandwidth than you thought you would be?
Often network providers are happy to increase the bandwidth limits on a contract but make it challenging to reduce those limits. Yet, with the power of real-time analytics and intelligent optimisation, an enterprise might find that its network usage is not what they thought it would be pre-transformation. Will providers be happy to renegotiate that limit down, even if it means a loss of ongoing revenue?
Many providers are talking about the positive benefits that network visibility will bring to their network, but this is one of the knock-on factors that the industry has yet to reconcile itself too. With increased visibility and automation, there are increased capabilities for reactive billing based on usage rather than large buffer amounts designed to cope with the surge. It will be interesting to see how vendors react and what enterprises demand as they see the full picture.
Looking for more thoughts from Cavell in the Networking space? You can download all recordings from our recent Cloud Networking Summit for free on our analysis page, via the button below.