Reinvention of The Communications Infrastructure Universe

Posted · Add Comment

Apparently I’m on some unintentional quest to piss just about everyone off in this world. What I mean is, last week I mentioned that I thought the network operators were going the way of the dinosaur and so now I need to see this through… have to go all the way.

A little background: I have been thinking this recurring idea/daydream for greater than a decade. 315f36a73435915e6d95ba38eb461d0a

It comes to me every time I step into a data center, Network Operations Center, Central Office, lab, cell site or other place where networking goes on. I love this analogy, but consulting in the networking industry is much like the movie Groundhog Day… again..data-center-pic1

Conversations I have with the engineering staff is generally the same globally…So according to your executives, you would like to deploy this and that service or feature, but this time you want to incorporate some amazing technology or cloud capability to cut costs and reduce the schedule. What is your current plan?

I get not possible/reasonablevariations of:

1) groundhog_day_minimal_movie_poster_by_bnxtd-d54wlrzOur primary vendor takes care of everything

2) We don’t have the schedule to do what management asked for due to some other priority. We just need to do the minimum as quickly as possible.

3) I don’t know what they are thinking, our vendors won’t let us do that.

4) We can’t do that, our roadmap … {either does NOT have THAT} {or has this in Quarter Z}groundhog_day_minimal_movie_poster_by_bnxtd-d54wlrz

In summary, most network operators have been driven by their primary network infrastructure supplier and are mostly hesitant to fully embrace independence.

Add chaos: there’s a spectrum of capabilities by these suppliers. Cisco, for example, tries to do EVERYTHING themselves, whereas some of the former telecom vendors such as Nokia, Ericsson etc that have been shedding people/services/capabilities since their peak in the late 1990’s to early 2000’s only provide their own equipment or staff for a few verticals and partner/resell everything else. So depending on the ask, the network operators get a highly variable quality of support.

The results are always the same. Tons of extra CAPEX/OPEX is burned for more or less no really good reason other than that’s the way it’s been done in the past. You always end up needing another upgrade, a new server, more development etc… Truth is the marketplace as is really doesn’t have any incentive to wind it all down and scale it all back. There’s jobs and business at stake.

Naturally this situation creates a multi-billion dollar services industry for staff augmentation/consultants etc.

computer+head+alternate+reality

The technology, a.k.a the state of the art, for the moment ignoring the changed world the smartphone app economy threw down, is being driven by open source and startup culture, and this is presenting a completely alternative reality…and that’s why I have ‘the daydream’.

Today’s technology gives a level of automation and efficiency to consolidate a lot of the current infrastructure running your network and/or services. It’s not hard to see the very automated, highly efficient future. For example: My car (with Navigation, auto cruise control and self parking) or Uber->Self Driving cars. Not hard to see. This is what a market disruptor will do for infrastructure. You may or may not see it coming.

I mean, there’s a new way to do almost everything in this market. None of these are dependent on a specific vendor, and are directly or contextually probably closer to free than you would believe. Here are some examples, top 10 style:

ITEM OLD NEW
1 Design your X by hand, lots of trial and error etc…using super expensive specialized tool requiring lots of $$$ and feeding Utilize a cloud service, crowd source, general purpose tools (ex: Google Maps)
2 Weeks/months of RFP process, vendor trials, etc… server blades for function X Software Defined Networking /Network Function Virtualization, OpenFlow eliminates custom hardware and allows software to perform most any function/service on demand
3 Deploy feature X on customized platform, or maintain legacy platform Move to generic hardware through virtualization like Dockers.
4 Optimizing your network with hours and hours of trial and error Self Organizing Networking (Plug and play)
5 Place network appliances everywhere and use Deep Packet Inspections (DPI) etc… Built in capability to DPI in new infrastructure and/or sensor fusion
6 Throwing everything onto a dashboard, (painstakingly) custom developed by hand for your situation. Probably runs on dedicated/custom hardware in your business. Cloud based 3rd generation monitoring and reporting
7 Using big iron such as SAP, Oracle etc..packages for optimizing your business with the associated hours and hours of trial and error Big data analysis.
8 Business Process Outsourcing (BPO) Platform/Software as a Service (PaaS/SaaS) like Amazon Web Services (AWS)
9 Workers Robots
10 Management Artificial Intelligence

There’s a lot of stuff out there in the datacenters, offices or wiring closets of the world but this list makes it clear that there’s a disruption speeding our way. For example, a wireless operator could today replace their network core, built blade by blade by Ericsson,  Nokia or Huawei and the associated care takers and logistics, with generic server blades either in house or in the cloud and software that could be pennies on those original dollars or free. If there is a major traffic changing event, a reconfiguration of this network could happen in milliseconds- no weeks/months of project managers planning cut overs and maintenance windows.

What’s driving this? Today it’s economics.

You can choose to continue to pay millions per application at each site to provide some service or because your infrastructure sales team and you have had a long relationship or spend on the level of 100 thousands to get equivalent or more capacity to do the same or more than you can do today with far fewer headcount. Oh yeah, there’s a new X that you need to do monthly. Your OEM can schedule the upgrade process or with the new way you can just instantly react….Also for the future, don’t underestimate the exacerbation of these economics that Internet of Everything brings.

If I were Cisco, Juniper, Ericsson, Huawei etc…etc….. I would be terrified that my existing business was getting torpedoed and race to plug law_of_disruptiona long term agreement for services or subscription into every customer nook and cranny possible.

 

Last week I wondered from a service provider’s point of view, who would I rather be, Orange, Verizon etc… or Amazon Web Services?

This madness is going to end. I’m daydreaming…and it’s the same question…who would I rather be: Huawei or a company like Twilio? There’s many more of those Twilio like companies popping up seemingly hourly.

We take care of the messy telecom hardware and expose a globally available cloud API that developers can interact with to build intelligent and complex communications systems. As your app’s usage scales up or down, Twilio automatically scales with you. You only pay for what you use – no contracts, no shenanigans.

https://www.twilio.com/

twillo_logo

Ok, so all the money you have invested in infrastructure… uh, ok, no need to remind you… Looking at the positive side: let’s go green and recycle your datacenters…I have connections with that too.