& cplSiteName &
Comments
Threaded  |  Newest First  |  Oldest First        ADD A COMMENT
brooks7
brooks7
5/1/2017 | 10:40:27 AM
A long reply
Apologies in advance, but your article posted so clearly why this is not working and I thought it deserved a thoughful and complete reply.

First: your commentary on why service providers are interested in Virtualization:
 - They want to reduce costs
 - They want to participate in new services

Those objectives are orthangonal to each other.  You can do one or the other or both and they don't provide momentum to the other.  I think this is the single biggest problem.  The whole standardization/specification cycle is about cost reduction by the introduction of multiple vendors who will price their products to drive themselves out of business to win at a large Service Provider.  The new services occur in an environment that moves so fast that the standardization will not work for them.  It takes too long to create this environment and this environment gets bypassed by fast movers.  Today's world is a perfect example of this.  IT service providers build products in Data Centers and using off-the-shelf choices.  By doing so, they have made Service Providers a Business Commodity.  The number of SPs that bring bandwidth to DCs is staggering.  By doing this, the IT SPs have made the network inside the DC very cheap and very fast and the connectivity choices many.  This move is being done and accelerated without participation by the SPs.  In fact, the participation that they did have has been largely sold off.

Second:  Your notion of a common API and essentially the demphasis on Open Source

Open Source is completely misunderstood.  When you are building a new IT service (like when I ran a SaaS operation) we looked at what OS we would use as part of the product.  The goal was to eliminate a number of software developers to get the product to market.  Why write when you can download?  Service Providers have been Systems Integrators.  OS does not work that way.  You need to provide the linkage between the pieces of OS UNLESS one project was designed to work with another exclusively.  There is no incentive for anyone to make an Open Source product that they pay people to write and then deliver to a telco for free - UNLESS there is consulting available.  I think you should instead look at loosely coupled packages that are connected through standard IP technology for what you are trying to do here.  This idea that a set of JSON objects or XML files are going to be completely plug standard and move instantaneously to meet the new services demands is just not viable.  The question you should ask yourselves is:  How many brand new services can be introduced per day by a SP?

Third:  A new app store

There are 2 falacies here.  First, that there is something special about a SP that lends itself to hosting an App Store.  Application authors want as wide a distribution as possible.  This means working to the lowest common denominator of connectivity.  Specialness is directly opposed to that.  If the SP is not special, than Google Play Store and iTunes will always defeat a SP.  The SP will be the Windows store equivalent (for mobile devices anyway) a complete afterthought.  Second, that any consumer of the app store will choose it first.  Why?  What will be compelling to make users move away from their existing choice of 3 app stores to choose one of the 100 new ones that will be created.

Fourth:  The wrong place to look

Sorry but you are looking in the wrong place and talking to the wrong people.  If virtualization is going to succeed than it needs to work for 2 groups:  Customers and Vendors.  What problems are you solving for the Customer that are not being solved today?  How are vendors going to make more profit by working with Service Providers virtualizing?  I think those are the key groups.  Talking to Service Providers about these is an utter waste of time.  What they have said (and I am about to paraphraze your article) is they want to make all the money off of Angry Birds while spending nothing and taking no risk.  Well, good luck with that. 

seven
Steve Saunders
Steve Saunders
5/1/2017 | 2:09:42 PM
Re: A long reply
Thanks Seven - 

Re: the vital groups. Yes, i agree. nothing happens without the vendors. We did actually roster ubiquitous support from all the NFVi incumbents for NIA last year, which surprised me (perhaps i shouldn't admit i was surprised?!?!)

Huawei, ADVA, Cisco, ZTE, Ericsson, Nokia - all in for the plan. Service providers have been harder to corale, but they are coming in now. 

My concept of a telecom app sote shouldn't be taken too literally, you know. It's more of a metaphor for interoperabile and viable commerical virtualized services! 

As always, thanks for the smart feedback... 

Steve 

 

 

 

 
brooks7
brooks7
5/1/2017 | 5:31:36 PM
Re: A long reply
Steve,

The problem with the telecom world is that the vendors are doing what the telcos say....even if it is bad for them.  Its okay with Huawei, because they are in the business of driving everyone else out of business.  But what I want to know is NOT if they are all in.  I expect that.  What I want to know how they are planning to raise prices with virtualization.  And if not raise prices, how they are expecting to raise profits.  That is the question that they need to be asking in the executive suite.  The telcos are peferctly happy for the vendors to price themselves out of existence...until they do anyway.  I am completely good if they say, we are terminating all of our Hardware Engineers tomorrow.  That is one way to raise profitability, but I hope you see where I am heading.

seven

 
brooks7
brooks7
5/2/2017 | 10:39:21 AM
Re: A long reply
By the way, by your definition I like the Telecom App store idea.  I want to throw out some thoughts about some boundaries around the services....

1 - Services have to save end customers money.  There are occasions where services can add to customer revenues, but that means a study of the niche and then being able to develop services aligned with that niche.  Not impossible, but I would think - second pass.

2 - Bandwidth Services and Transmission Services have to be ancillary to the problem solved.  The competition is IT services connected over by MPLS, the Internet, or a Hybrid of both.  Those services will be developed quickly and deployed ubitiquitously.  If you tie your special network to the service, you are likely to serve only a small minority. 

3 - Consider an Ecosystem play as the answer.  Instead of trying to compete with these IT services, consider instead a way of making them easier to develop and deploy or more secure or easier to access.  Revenue would come from increased network usage not higher value of a bit per second.

4 - Operational considerations at the customer should drive approaches.  IT folks are hard to hire and are expensive.  How can you reduce the need for end customers to hire more folks?  Be a partner and not a vendor.

5 - Speed is of the essence.  In today's world, switching costs are high and are avoided as a general rule.  That means that new applications are a land grab game right up front.  So unless you are considering a "Levi Strauss" model (see item 3), you need to be first or near first to market.  If you are not the value has been sucked out by the time you get there.

seven

PS - Telecom equipment vendors...read #5...it is the way that Service Providers have been buying for over 10 years.  You either win up front or you don't win.  Quit throwing good money after failed product lines with the old Product Management/Marketing Hockey Sticks (We are at the inflection point right now!).

 
Steve Saunders
Steve Saunders
5/2/2017 | 4:40:15 PM
Re: A long reply
"What I want to know how they are planning to raise prices with virtualization.  And if not raise prices, how they are expecting to raise profits. "

 

This is the exact problem Cisco is wrestling with with its software licensing right now, yes. 
tojofay
tojofay
5/2/2017 | 6:54:26 PM
Re: A long reply
some are leading the way: https://www.infinera.com/eurofiber-scales-service-offerings-with-end-to-end-infinera-network/
tpohara
tpohara
5/3/2017 | 12:06:12 PM
Excellent Article
I could not agree more with your article. We spent years re-engineering our core product(s) to address this very concept within the industry. 
danielcawrey
danielcawrey
5/4/2017 | 11:49:36 AM
Re: Excellent Article
Interesting read. As a technologist I do find virtualization interesting. However, I can see how constantly talking about it and its implementations can get old. The real business benefits of this are really what the conversation should be about these days. 
kq4ym
kq4ym
5/13/2017 | 2:19:46 PM
Re: Excellent Article
As noted, "standards, certification and interoperability" are the keys to making it all work. And it will be interesting to watch as APIs are developed and especially to see how the standardization of certification fees may work out to get a level playing field here too.
Mgagnon1956
Mgagnon1956
5/3/2017 | 3:10:41 PM
Telecom App Store - Its all about the Business Applications
Steve,

Great perspective on a vexxing challenge for Telco's. The idea of a Telco app store is timely.

What the Telco app store needs are business applications linked specific network virtualization enablers. Why one NFV service versus another? What underlying business benefit is dervied from linking Business App "A" to NFV Service "B" or "C". Does it matter and why? 

At Nubo Software our platform allows Enterprises to de-couple mobile apps, data and security from a physical mobile device and shift them to a virtualized mobile device running on a Telco Cloud. Virtualized apps such as Office 365, Outlook, SFDC & others are streamed back to physical devices over Telco connectivty fabrics that would benefit enourmously from NFV services to power a specific Qos user experience for the app. 

"Zero" data is transmitted to or resides on the device making the mobile session perfectly secure. NFV services allow high performance bandwidth & latency to be "dialed up" on demand allowing an amazing virtualized app experiences at Telco Cloud scale! 

We think mobile app virtualization (MAV) coupled with NFV services and MEC/5G, offer Telco's a tremendous opportunity to render business apps and NFV services from the same Telco app store. 

Your thoughts?

 

 

 
RMersh
RMersh
5/5/2017 | 3:36:43 PM
Standards, Interop, OS, migration and the business end of NFV
So where do I start.....

First, I do appreciate the thought that has gone into both of the articles you have posted on the relationship between OS and standards, and what operators should focus on to make a success of NFV and SDN. You have made a number of good points, although I think some issues are missing and some points are made a little aggressively. The problems you identify are fair game, but we do need to think hard on what can make a difference in the resolving them.

So, on the issues you identify, I agree that:

- virtualization is not yet delivering on it's promise

- open source development has not been successfully married to the best practices of standards

- interoperability should be a key consideration in virtualized products and services

- the introduction of virtualization is slower than some commentators predicted

- there has been too much hype and not enough reality

- standardised APIs are critical 

- virtualized solutions may be over-engineered

- operators are really looking for new services / revenues

- operators want an open market place for virtualized products

You were a little harsh on some actors in the industry. As I'm from the standards world, you may be surprised that I want to defend our colleagues in the OS community. They are not as 'flaky' as you insinuate, we were really quite inspired about how quickly and how concrete they have been in developing solutions, ones that show how we could develop standards at the Broadband Forum in completely new ways. OS is distruptive because it is producing more than just powerpoint presentations.

Some points I expected you to make though, were absent. For the broadband community, we are keenly aware of issues like co-existence between traditional networks and programmable / virtualized functions, migration paths and best practice, the requirements of the smaller operators (who are critical to the market place) and potential regulatory questions.

I think you are right to emphasize the benefits of standards - industry consensus (market development), promoting conformance, 'safe harbor' and the discipline of the standards process. What is needed though, to gain the advantage of NFV and SDN, is to use the output of OS projects and the commercial software of the vendors in a standardised framework where you can prove new services, test them and develop best practice for deployment. We just announced a new initiative for Open Broadband on exactly this:https://www.broadband-forum.org/news/download/pressreleeases/2017/PR07_BBF_China_SDNFV_FINAL.pdf

Lastly, you are right to emphasize APIs, where some of your predictions / indications may already be happening.

Some providers who are already engaged at the BBF have contributed an industry level Telemetry API for application and cloud portals that basically form a migration path using legacy and cloud telemetry models.This came to the BBF so as to ensure the Residential Gateway specifications will follow through, and support an operator consensus on what the industry "managed resource model" will be.

I wish I could be at BCE to take up your offer to discuss this further, but unfortunately I am double booked.

 
Tomasz Stepniak
Tomasz Stepniak
5/7/2017 | 12:32:06 PM
Long way to go...
I can see three major risks:

1. Security

2. Security

3. Vendor push back as the vendor-lock is very convenient to reduce all operational risks. 
Solution63465
Solution63465
5/19/2017 | 4:24:02 PM
Metaphors
I really liked the article and whatever points were being made-

Except for one- Is is necessary to use the term "boner" in the writing?

Perhaps there are some readers who cannot relate to that experience.

Way to alienate approx 50% of the working population...no?

March on SYS-STEMMERS!

 

 


Featured Video
Upcoming Live Events
October 1-2, 2019, New Orleans, Louisiana
October 10, 2019, New York, New York
October 22, 2019, Los Angeles, CA
November 5, 2019, London, England
November 7, 2019, London, UK
November 14, 2019, Maritim Hotel, Berlin
December 3, 2019, New York, New York
December 3-5, 2019, Vienna, Austria
March 16-18, 2020, Embassy Suites, Denver, Colorado
May 18-20, 2020, Irving Convention Center, Dallas, TX
All Upcoming Live Events
Partner Perspectives - content from our sponsors
Edge Computing, the Next Great IT Revolution
By Rajesh Gadiyar, Vice President & CTO, Network & Custom Logic Group, Intel Corp
Innovations in Home Media Terminals for the Upcoming 5G Era
By Tang Wei, Vice President, ZTE Corporation
All Partner Perspectives