Much of the focus around virtualization right now is on getting the right approach in place to manage and orchestrate software-defined networks and virtualized functions. But there is another pressing challenge to this critical network transformation that the telecom industry needs to address quickly and as a whole.
There needs to be an organized effort to redefine net neutrality in the virtualization era in a way that allows the industry to move forward with creating network flexibility and efficient use of resources, without fear that regulations are going to limit the ways these can be commercially applied.
I'm not talking in any way about a draconian move that undermines the core principles of open access to the Internet, or seeks to undermine innovation. I think the telecom industry's interests are quite the opposite -- but that is not the public perception. While telecom network operators are facing the very real threat of being made irrelevant by web-scale content and service providers, they are still publicly perceived as the Internet's all powerful bullies, looking to suck the life out of Internet startups and innovators.
That's why there needs to be a positive discussion around the benefits of things such as network slicing -- applying the specific resources required by an application -- and other application-specific impacts on the network, to those beyond the telecom inner circle. Some of this discussion must involve those who stand to benefit, who should be enlisted as allies. In this case, I would specifically engage apps developers, smart city leaders, higher education standard bearers and companies engaged in Internet of Things efforts in partnership with telecom.
There has been relatively little discussion among US players about the impact of net neutrality rules on NFV deployments -- that doesn't mean there isn't the chance that the same forces who rallied public opinion to get new rules imposed at the federal level last year won't rise up again as virtualization becomes more real.
The best way to forestall such action is developing a realistic and honest picture of how the competitive landscape can and should evolve in the virtualization era. Telecom can tell its own story now or wait and let someone do it for them later.
Re: Open Access Absolutely. Especially if both of these companies really want their Millennial-targeted services to resonate with younger customers. These younger users seem notably more savvy and less tolerant of this kind of legacy protectionism, which I find in and of itself pretty fascinating.
Re: Open Access Correctly noted, the changing political winds not to mention the public's own take on the industry can have big changes happen at least in the distant future. It's still true the telecoms are "publicly perceived as the Internet's all powerful bullies," and that probably will have to change.
This is the sensibility that is missing from everyone discussion: "The whole notion that we are going to build more intelligence into a network instead of into applications is just backwards".
My point about this is that if we have to stretch ourselves and put something in the network domain instead of the application domain (otherwise you can just build the appropriate controls into the app), then the whole thing is more easily solved in the application domain. The whole notion that we are going to build more intelligence into a network instead of into applications is just backwards.
It is a really cool idea from a carrier revenue perspective. Hey I can charge more for specific bits per second AND do it with virtual devices. More money no new capex. The problem is that it doesn't work for the application developer. Think about it from your fake case. Unless you can signal lots and lots of carriers and the price to do so is really cheap then why is it a good idea for your application? If you require that, then you probably have a business model that won't work.
seven
Edit: Application domain control = something like what happens with Youtube streaming when it encounters congestion.
Re: Open Access There's always congestion somewhere.
Yes, I'd love a congestion free service, however I think a provider may be unwilling to give me 5Gbps to my home for $20pm, or even better, for free.
Yes, I'm making an assumption that modern apps are designed for less-than-optimal bandwidth to cope with over-subscription. However we both know that if there is more bandwidth, developers will find a way to use it.
So, a suitable method is needed to manage the situation.
I'm assuming people want to assign the highest network preference to that video call about the new baby, rather than the streaming music videos for the kids.
From an end-user point-of-view, it's a dial that prefers this session over that session. From a service provider point-of-view, that hint might allow them to tag/constrain traffic somewhere upstream for the user and provide a better quality of service for that video call, without inspecting that call. Yes, this assumes the congestion point can be influenced by the retail service providers.
Geez, I'm trying to defend an imaginary use case. I'm just trying to find a situation where" more bandwidth" isn't the simplest answer.
Re: Open Access "I can imagine a user level QoS app that would allow a residential end-user to signal to the network that a certain class/source/type of traffic should be preferred."
That violates net neutrality so in fact it will never be developed. On top of that, it would only work if every carrier involved allowed the traffic to be prioritized across the entire path. Otherwise you would not be able to predict what happens with congestion. But let me ignore all this for the moment.
And I need the exact functionality of the app that requires prioritization. Not just, hey I want my traffic prioritized over everything on the Internet. Given that real-time video games and video streaming don't require this priorization, I declare that no RESIDENTIAL application requires it. Even if that is a way to solve a problem an easier solution is add more bandwidth to the network. It accomplishes the same job. Prioritization ONLY matters under congestion. If there is no congestion, then all streams run at the maximum possible rate.
that's quite a challenge you've set for our fellow readers!
I can imagine a user level QoS app that would allow a residential end-user to signal to the network that a certain class/source/type of traffic should be preferred.
I'm thinking of a screen that show real-time network traffic on the users gateway, then a network-authenticated user (eg, the account owner, or parent) has a dial, or levers that would visually help them prioritise the outbound traffic. A user level app, if network slicing is in place, could allow the residential user to signal to the upstream devices so inbound traffic could also be influenced.
In terms of revenue... hmm that's trickier, advertising driven? Subscription to maintain the upstream controls? Consume tokens every time you make a change?
I would like an example of an application that would get developed that would require special processing in the network. This application MUST be able to be deployed without this (otherwise how would it get developed). And that this MUST be more flexible than available to anyone, anywhere with an Internet connection. This application MUST apply to residential customers.
The reason we are talking past each other is this. You are thinking like a networking person not an application developer. The whole point of this is to make it better for developers in your message.
Q1: How will I as an application developer make more revenue by using the network resources?
Q2: How will I as an application developer spend less to operate my application?
Now there are plenty of examples of these - CDN's. But think about how basic this is. Are they saying they are going to dynamically create and sell CDNs based on virtualization? Imagine the billing nightmare that would be. Johnny over here started watching a video, so I cranked up this CDN. Hey, wait - 1 guy streaming and you are charging me for extending the CDN?
Carriers want to charge more for special bandwidth. They have been at trying to do so for years. They can't right now because of regulation. But I would argue that they can't even without regulation because it is a terrible business idea. The carriers will spend bucketloads of money on equipment and software that nobody will ever use. This is because they look at things by estimating a way to charge more for what they have without understanding the value proposition to the application developer who would buy it.
Re: Open Access I feel like we are talking around each other. Your explanations don't fit what I am discussing. This isn't about Uber, and it definitely is about making applications work better - or rather making networks better ble to support applications in a way that is more flexible.
Make the right resources available means that it is special. Think about it. How many special network resources are required to surf the web?
That is what apps and app hosters are DESIGNED for - best effort Internet. They are designed to require no special latency or security requirements for RESIDENTIAL access. Business Access is different. Net Neutrality only applies to Residential access. If you design a service mandates the need of special latency, then you can ONLY send it over specific networks. Therefore, it is useless to create it. So, sit in the chair of the makers of Angry Birds. Would you really develop an app that only runs on Telenor's network but can't work on China Telecom?
Take Uber. Imagine if they built their app to only work on Verizon Wireless. Or any single carrier. That is why none of this works. Not because it is not a nice idea for carriers. It is a terrible idea for app designers. So they are right if you have to have an app that requires super specific network function, then it won't get developed. But I have the existence proof of millions of apps that exist, across uncounted number of functions and NONE of them require specific networks. Even Netflix.
Now why are business applications possibly different? You have a restricted population of users and can probably guarantee a significant percentage of the audience is on specific networks and devices (Note even this is going away slowly with BYOD).
What this is all about is ways to generate extra revenue by carriers by encouraging application developers to need things from them. It has nothing to do with making actual applications work.
Many larger network operators are bringing software development in-house, developing their own orchestrators and other tools, and using more open source to get the results they want, without vendor dependence.
To save this item to your list of favorite Light Reading content so you can find it later in your Profile page, click the "Save It" button next to the item.
If you found this interesting or useful, please use the links to the services below to share it with other readers. You will need a free account with each service to share an item via that service.
(19) | 
Digg
Del.icio.us
Reddit
Tweet This