acronym heaven - the real future of 5G
When an executive from Nokia, of all companies, said 5G was as much about fibre as wireless, it was clear this was going to be different from previous mobile standards generations. 5G will not be driven by mobile broadband speeds as 4G was. If higher data rates and larger numbers of broadband devices are an mobile network …
It's purposeful obfuscation.
The mobile networks want to maintain control of 5G when the distributed nature of nodes plopped on t'internet mean it can be controlled by anyone who groks OpenStack. So that'll be the cloud suppliers like Amazon.
To be honest, all these different network function virtualization 'nuances' introduced by the protagonists are unimportant. To people used to the internet we know the power of gateways to join dissimilar segments together.
There's only one thing important to everyone and that's billing. And no-one can stop the inevitable where money is concerned.
Yes, the blockchain is where all billing (rating records) will go. Monolithic network operators are going the same way as banks. Totally unnecessary.
But I struggled on this article, glad to know other's did. I hate Acronyms at the best of times too.
No sure who it was written for, but if you fully understand it, in terms of the software overview/hardware overview - you're a Telecom God, compared to me (and I feel I'm conversant in Electronics/RF Radio/Analog Circuits and Software Development/Virtualisation,Cloud interoperability).
Even then at times, it can often feel too much like a jack of all trades, master of none, to get some sort of overview of what 5G entails, boy this can be complex.
The best way I've found to make sense of all this is, whenever I get a chance, is to talk to the field techs and engineers. Everybody seems to like to talk shop when they are out&about. Then again I talk to everyone, yes, even the squirrels. [That's a stock joke in the US Navy about nuclear types. It's true.]
Elephant 1. 5G is "more" compared to 4G/LTE, but more of what? Speed? Throughput? Users already have quite a lot of that, and aren't inclined to pay more. To put it another way, there is no killer app that will drive 5G adoption, and when it arrives, it's unlikely that the telcos will be the ones to invent it. So then you have what you might call Netflix-on-steroids, where the money is all flowing to the app creator over dumb pipes that are not being monetized. No telco wants to spend billions building out 5G for that.
Elephant 2. 5G invents some new ideas such as mobile edge computing where work can be done for a user right at the edge of the network. Say, rendering for your augmented-reality world done there so that your personal AR system doesn't need a ton of local horsepower. Not my favorite idea, but still. But here's the thing. This edge computing thing needs to be able to move around, e.g. when you drive down the road and switch from cell to cell, you don't want your AR display to disappear. So now the compute infrastructure must be able to do a sort of superfast vMotion-type of rehoming. Think OpenStack or VMs can do that, in under 5ms (the standard)? Think again. Containers? Maybe...
Elephant 3. Hate to say it, but re: ONAP, OSM and so on, the emperor has no clothes. Teams of tailors are busy sewing, but we're not much past the Borat swimsuit level of coverage. There's still a lot of Grand Unified System stuff going on here while the small mammals of Ansible and Chef are doing little things well. Which method will win out is TBD. One of them needs to win quickly so that devs who are unsure which to commit to can plunge in.
Elephant 4. Container networking. The mobile network is a complex beast for good, sound technical reasons. IPSec tunnels, encryption, traffic shaping, policy, AAA, it's got it all. Good luck getting all that going in containers.
In short: lots to do. And not much money at the end of the tunnel.
Killer app? Watch for virtual reality, augmented reality and possibly holographic projection. It's coming.
Also, as with all new tech, the rich early adopters pay the premium while costs for incumbent technology drop for the rest of us.
Dialup to broadband.
Cassette/floppy to CD
CD to DVD, DVD to Blu-Ray
CRT to LCD
LCD to LED
HDD to SSD (happening right now)
I thought it was a good article but I do have a few comments that may be of interest to the telco geeks.
1) T-MANO in an implementation of OSM, based on HPE’s NFV Director, not a rival standard.
2) Open Source MANO is just as open source as Linux's ONAP. That ETSI has embraced open source (albeit just for this project) is revolutionary but it is real.
3) Verizon is not on the fence. They joined OSM in February and have said they have no plans to use ECOMP (and presumably ONAP).
4) OpenStack is not an alternative to ETSI MANO. It is an option for the Virtualised Infrastructure Manager component (alongside VMWare). OpenStack’s project Tacker is working on a generic VNF Manager and NFV Orchestrator but this isn’t the core OpenStack application and I don’t think it is ready for prime time just yet.
5) ETSI is not resisting a model-based approach. The Network Functions Virtualisation – White Paper #3 states NFV “requires a new set of management and orchestration functions and creates new dependencies between them, thus requiring interoperable standardised interfaces, common information models, and the mapping of such information models to data models.” Sounds pretty model-based to me.
If this sort of stuff interests you can check out my report: Next-Gen OSS for Hybrid Virtualized Networks
or get your fill of telco acronyms at www.lightreading.com
Biting the hand that feeds IT © 1998–2019