Addressing Solvency of Open source production and adoption models for TMT and Telco industry

An architect’s perspective

Source: http://www.cncf.io

Technology, Media and Telecom industry known as TMT by commoners is going through major transformation programs globally. One of the prime recipes of this revolution are their vigor and participation in Open source. However, over years of experience in Open source revealed some key points to us, For example most of industry believes opensource is

• Open (without defining what is Open)
• Cheaper (with our a valid Business case and TCO working)
• Simple to deploy and use (with out analyzing ecosystem and interworking).

On one hand I believe Opensource is a bandwagon every one wants to sit but is also a wagon no one want to drive at least in a commercial and production environment. Therefore, it is very important to share my views on this .

Understanding the Story of TMT Industry at a Glance 

Pictures are always an easy way to summarize and let me share some very useful insights from lumina networks  in recent #ONS Days Melbourne showing Opensource will address solve business issues around . Thanks for IldikoVancs for sharing this summary .

• Automate any network with any form factor
• Resource optimization and utilization
• Offer NaaS and Slicing to Verticals most important 

Attracting business to Opensource is the most important challenge for Architects , Will building a best of breed staff fulfills TCO and CXO transformation objectives ? 

Source: @lunminanetworks

https://twitter.com/IldikoVancsa/status/1193684616869113856/

Let us see now results of applying story to practice through Impacts of Open source Glory in TMT industry

Use of open source in Telco has definitely a long-term value but because still as a community we have failed to transfer Open source frameworks to MVP of vendor products, some thing that can drive and navigate our Sourcing, RFX and vendor selection process. I think this was something that was supposed to be done at the outset of this journey but what really happened. I think some smart marketing crooks sold dreams to Telco’s that they will become future Google and hyper scales which was never the purpose and direction of a service industry like ours.

The long term repercussions of this was that still down the road of almost a decade we found ourselves surrounded by vendors that come with a proprietary and engineered solutions  that they say is aligned with Opensource like ONAP , OPENSTACK,K8S etc but the sad story reveals its just the concept or in SDK terms the front end that is something like Open while all backend and software forking is proprietary .

So despite I lead my company in many of these initiatives I am still circumspect of the approach we have taken or at least we should agree that we need to really now think about value quantification to put business metrics in perspective and to ponder how to really make these initiatives work.

Grow adoption through power of synergy

Openstack is one of the most successful example of Opensource adoption and summing our experiences in Cloud may support to build the best prospective in other domains, may it be technical issues , licensing models or operators and TMT’s transformation journeys . The wide adoption of Openstack in IT , application, Telco and verticals proved that opensource frameworks that can serve a wide use cases not limited to particular industry is the best way to proceed. It has really enabled to adjoin industries which were traditional on a minimum of an Isle’s distance.

However balancing an opensource adoption in a balanced way with out skewness towards a particular segment is still a big challenge, it comes a life line when we consider each industry requires unique characteristics and building something for commoners cannot fulfill business needs.  This is what leads industry to collaboration or which in business or RFX we name as partnership and ecosystems.

Sharing and not swallowing

As Johanne Mayer, Director MayerConsult and TM Forum Distinguished Fellow at the Layer123 Network Transformation Congress in San Jose the Telco’s in 5G era can not find a solution that comes from a single vendor. It simply is not possible as more and more verticals will come in to the picture. The success of such initiatives will depend more on working together then to get community outputs and build end to end solutions and products that simply are not aligned with open standards and concept of mixing best of the breed approach.

Integrating the Open source solutions

Open source is a journey of software and idea that future the solution will come from a hefty number of vendors. However, it is still a reality that at least till the decade Telco’s will run the hybrid networks including Cloud , PNF , VNF etc . So we need to solve these issues on API , standards and Fitness for use at the outset .

Also we need to see use of open source in customer X can be totally different than the use in vendor Y because the end product and business case will require some solution tailoring. The Personalization of each customer solution is very important.

RFX of Open source solutions

One of the bigger pain points in Open source solution adoption is to do with sourcing and RFX process. It is easier part for Telco to give requirements of an open solution but fact is which vendor will take end to end responsibility of third party in such multi-vendor offering.

Then how an Operator can be sure the vendor is not pushing its solution whereas ideal solution must be the Pull solution where based on requirements each component should be selected and then integrator to work on modalities of integrating them all . It is easier said than done however failure to adopt this direction means we find an integrator that can only work with certain vendors or at least can only take project if certain vendors on certain later are there. This is a frightening situation that will impede to slow adoption and commercialization of open source technology. If we address these points not only on technology level but also process and sourcing model then I am sure adoption will be as freely distributed as the Windows of today.

Its Services not the solutions

Some time ago I was reading a blog from NetApp  with a sentence that caught me in the moment

“You Can’t Change Your DNA – EMC Thinks Mainframe; Dell Thinks PC; NetApp Thinks Open Systems”

That is so true for the TMT and specifically the Telco industry , we must not be lured by story of openness and innovation that we forget our purpose. I think key to our survival in new era of digital transformation is fail fast and to deliver services .  Focusing on services and not on solution is so vital for Long term sustainability.

Source: https://www.dreamstime.com/

 When ROI of Open source solutions will realize

Although painful but reality is that no CSP unless the giants with big R&D arms have achieved real ROI or a fair visibility into it at least till now. One practical result of this is slow down strategy by many CSP’s and Telco adopted , however slowdown simply solves issue but delays it only .

Lately I have seen many Telco’s in EMEA to solve issue by breaking stack from the top or North side then to fix it from ground (NFVI upwards) however it a child’s cry unless we  at least agree and define clearly what is the definition of automation

The risk of industry failure to solve this issue is all major Telco’s are considering proprietary solutions of orchestration and automation to have a feel of something practical that can do something realistic on automation.

Then for real ROI we must agree on something which will finally scale , as an example spending X years while building two Data centers NFV stack we find once we will scale to lets say 10 the advantages of scale is simply not there. The issue of Open source introduction at the Edge will make this discussion more prevalent as simply an Operator can not afford to go to Edge with high setup cost, had open source solution not fix this side will obviously result in the slowdown of Edge as a whole.

Finally, just solving OSM and ONAP architectures with out ensuring vendors MVP align with it and that finally automation stacks can be built by combing best of breed from different vendors is and will remain a dream unless definition , reference architecture , reference implementation and testing/validation is not agreed and vendor’s are not enforced to apply this across their products . Although latest work on hackathons (LFN developer and testing forum)  have tried to address the same concern however still author believes it is not production ready . This is how LFN defines it and I think this is where standardization and SDO’s need catch vendors to align their MVP with architecture frameworks and standard .

“Co-hosted with GSMA, this LFN Developer & Testing Forum brings together developers across ONAP, CNTT, and OPNFV, with a special focus on VNF compliance and verification testing. As the principal event for LFN open source projects, the technical gathering provides a platform for community members to collaborate, plan future releases, conduct a plugfest, and perform interoperability, compliance, and integration testing across the full network stack, including NFVI, MANO, and VNFs”

 

 Future lies at the intersection of our path

With so many things happening in the industry around both open source development and testing and its commercialization, it is not surprising that many carriers especially those outside U.S and EU to follow a more conservative strategy

However my takeaway to CXO will be to ask technical team focus on bigger picture to know how and when the proprietary automation or Open source solution will converge to real open solution opening choices and offerings for the service providers , had this not addressed the Open spruce journey will remain a glory which every body want to talk about but a road no business wish to take finally .

 

My final thoughts on this topic is actually best described by a session on O-RAN by Orange in ONS Europe 2019 which is

“The Price to pay for Open source greater flexibility, innovation and openness is the complexity of test and integration” .

So finally, all will come to one line how to define MVP products, benchmark the solutions and about all develop a common test and integration model. It is clear that open source bring value but if we do not know how to deploy and fix the issue it becomes a nightmare nobody wish to keep. This is where SDO’s and Operator community needs to focus in 2020 . One such direction is the CNTT (Common NFVI Task Force) and CNCF TUG (Telecom User Group) which are expected to solve very issues highlighted above .

Currently we are working on the CNTT R2 which is expected to be a GA by JUNE 2020 . Similarly R2 of CNCF will combine the testing and validation of open source solutions to adress the issues that had been faced by TMT clients .

https://wiki.lfnetworking.org/display/LN/Common+NFVI+Telco+Task+Force+-+CNTT

Advertisement