Shifting Gateway paradigms with Virtualization
17 June, 2021
Virtualized networking has been a huge topic for years already. The concept of removing intelligence (i.e. software complexities) from networking equipment and running it in central servers has been used to make large networks cheaper and more efficient. Underlying much of this is a protocol called OpenFlow that the servers use to program various types of networking equipment that Operators have deployed in our neighborhoods (and beyond). More recently, many Operators have been looking at the same OpenFlow as a way of adding features to the gateways that we have in our homes for Internet access. This is done by integrating a software module called OpenVSwitch (OVS) into the gateway software.
Modern gateways are really communication powerhouses, providing a wide range of complex operations so that we can enjoy a (hopefully) fast, reliable and feature-packed Internet experience. Our Internet connections today do a pretty good job on the “fast and reliable” part. Connect rates have increased dramatically over the past years and the basic networking protocols needed for connectivity are very much commoditized and accelerated in silicon.
On the “feature-packed” front, though, things are not so simple. Gateways are growing in complexity to provide not only speed but all sorts of functions – or services – to improve our experience on the Internet. Services may be used to prevent children going to abusive web sites, for checking that elderly parents are ok, to help us improve Wi-Fi reception throughout our homes and lots more. In fact, lots, lots more.
Delivering these services to users is very slow and inefficient. It takes literally an Operator ages to get new functions available for their users. Debugging, upgrading and maintaining these services is a whole different story and an even larger headache. Standards bodies have defined mechanisms that somewhat alleviate this pain (with containers, USP, etc) but these are mechanisms for deploying services, not a complete solution for delivering them to end-users.
Our solution is the Inango Service Orchestration Platform. It is a software platform that runs on external servers and connects with gateways using OpenFlow. What our platform does is to allow Operators to deliver an unprecedented wide range of services to their subscribers via the gateways via a fully automated “orchestrator”. Inango not only deploys services – we apply user preferences for each, optimize traffic flows among them, only allow services to access permitted parts of the gateway system, keep extensive logs and alarms so Operators can check things are working properly. And we do more than that – if your gateways are out of memory, you can even deploy some of the services at your data center without losing any functionality. Operators are empowered to pick and choose which service they want and where it will best run for each individual user.
In short, Inango offers the Operators unparalleled Scale, Efficiency and Choice. Scale means the ability to grow a service portfolio almost unlimitedly. The ability to deploy services locally (on the gateway) or remotely (on a server) means memory sizes on gateways is no restriction. And having automated ways of maintaining the whole process means that Operators can actually manage a large portfolio of services. Efficiency means making services available to all users instantly, at the flick of a switch (whether they are deployed locally or remote). No intervention, no field upgrades of gateway software. And Choice means that both Operators and user may decide which type of service they wish to use on their own Internet connection. Things not working with one type of service? You can simply switch to another. It is like having a “Supermarket of Services” that you and pick and choose from, much like we do on our phones today. We can’t go through all details here, but if you’d like to hear more, please contact us and we’ll gladly give you a demo.
We’ll talk about the “how” part of OVS next blog.