Tuesday, September 29, 2015

SDN NFV in Telco 2015-2016 exec summary

2015 has been the year that SDN NFV efforts “got real” in telco networks. Past the first two years of enthusiasm that were marked by unrealistic vendors’ announcements and a flurry of participation in standards, open source and proof of concept projects. These efforts certainly put the technology on the map and have set operators and vendors priorities towards the exploration of the maturity, promises and limits of the technology.

One of the main problems with a revolutionary approach such SDN and/or NFV implementation is that it suggests a complete network overhaul to deliver its full benefits. In all likeliness, no network operator is able to operate fully these kind of changes in less than a 10 years' timescale, so what to do first?

The choice is difficult, since there are a few use cases that have seemed easy enough to roll out but deliver little short term benefits (vCPE, some routing and switching functions...); while the projects that should deliver the highest savings, the meaty ones, seem quite far from maturity (EPC, IMS, c-RAN...) in a multi-vendor elastic environment.

The problem is particularly difficult to solve because most of the value associated with virtualization of mobile networks in the short term is supposedly ties to capex and opex savings. The business case for saving based on new infrastructure introduction is difficult to make without compelling new revenues streams financing the architectural upgrade.

Islands of SDN or NFV implementations in a sea of legacy network elements is not going to generate much saving. It could arguably generate new revenue streams if these were used to launch new services, but today’s focus has been so far to emulate and translate physical function and networks into virtualized ones, with little effort in term of new service creation.

As a result, the business case to deploy SDN or NFV in a commercial network today is negative and likely to stay so for the next few years unless innovative services are launched. I expect the momentum to continue, though, since it will have to work and to deliver the expected savings for network operators to stand a chance to stay in business.

The other side of this coin is the service offering.  While flexibility, time to market and capacity to launch new services are always quoted as some of the benefits of network virtualization, it seems that many operators have given up on innovation and service creation. The examples of new services are few and far between and I would hope that these would be the object of more focused efforts.
OTT services explosion, combined with the progressive opacity of the traffic due to encryption conspire to make network planning extremely difficult. Peak traffic are unpredictable and increasing in frequency and magnitude, which makes the rationale for network capacity purchase based on dedicated appliances untenable from an economic standpoint.

Revenue stagnation is a given, with little in the way of new streams from new services such as VoLTE or M2M for a few years.

Cloud technology seems to be the key to the new OTT providers’ agility, but its implementation supposes a complete organization, structural and process upheaval that network operators are hesitating to implement without a firm business case.

SDN remains a reliable, mature technology for enterprise and IT cloud management and traffic switching but requires important efforts to adapt to telecommunication networks, mindsets and regulatory frameworks.

NFV is emerging as a key potential effect multiplier but, without a viable service orchestration framework is becoming a collection of large proprietary frameworks from legacy telecom equipment manufacturers or an endless suite of isolated virtual network functions with little coherent cohabitation model for a harmonious service delivery.

All is not lost though, with an ecosystem that is moving faster than any initiative in the telecom standards world, it feels that the vendors, operators can find the right recipe within a few iterations, unleashing a flexible, scalable elastic environment for the cost-effective  creation and management of tomorrow’s services.

There is certainly a race between the likes of AT&T, Telefonica, Deutsche Telekom on the service provider side and Affirmed Networks, ALU, Ericsson, Huawei and HP as some of the leading vendors to deliver on the promises of SDN and NFV in telco networks.

This report provides a review of the main trends that are pushing network operators to move towards simplified cost effective network architecture and the vendors’ strategies and roadmaps to address this disruption in their traditional architecture and revenue model.

Thursday, September 24, 2015

SDN-NFV in wireless 2015/2016 is released




As previously announced, I have been working on my new report "SDN-NFV in wireless 2015/2016" and I happy to announce its release.

The report features primary and secondary research on the state of SDN and NFV standards and open source, together with an analysis of the most advanced network operators and solutions vendors in the space.

You can download the table of contents here.







Released September 2015
130 pages

  • Operators strategy and deployments review: AT&T, China Unicom, Deutsche Telekom, EE, Telecom Italy, Telefonica, ...

  • Vendors strategy and roadmap review: Affirmed networks, ALU, Cisco, Ericsson, F5, HP, Huawei, Intel, Juniper, Oracle, Red Hat...

  • Drivers for SDN and NFV in telecom networks 
  • Public, private, hybrid, specialized clouds 
  • Review of SDN and NFV standards and open source initiatives
    • SDN 
      • Service chaining
      • Apache CloudStack, Microsoft Cloud OS, Red Hat, Citrix CloudPlatform, OpenStack, VMWare vCloud, 
      • SDN controllers (OpenDaylight, ONOS) 
      • SDN protocols (OpenFlow, NETCONF, ForCES, YANG...)
    • NFV 
      • ETSI ISG NFV 
      • OPNFV 
      • OpenMANO 
      • NFVRG 
      • MEF LSO 
      • Hypervisors: VMWare vs. KVM, vs Containers
  • How does it all fit together? 
  • Core and RAN networks NFV roadmap
Terms and conditions: message me at patrick.lopez@coreanalysis.ca

Thursday, September 10, 2015

What we can learn from ETSI ISG NFV PoCs

This post is extracted from my report SDN - NFV in Wireless.


Last year’s report had a complete review of all ETSI NFV proof of concepts, their participants, aim and achievements. This year, I propose a short statistical analysis of the 38 PoCs proposed to date. This analysis provides some interesting insights on where the NFV challenges stand today and who are the active participants in their resolution.
  • 21 service providers participate in 38 PoCs at ETSI NFV
  • 36% of service providers are in EMEA and responsible for 52% of trials, 41% in APAC, responsible for 25% of trials and 23% in North America, responsible for 23% of trials.


Out of 38 PoCs, only 31% have seen an active participation from one or several operators, the rest of the PoCs have seen operators take a back seat and either lend their name to the process (at least one operator must be involved for a PoC to be validated) or provide high level requirements and feedback. The most active operators have been Deutsche Telekom and NTT, but only on the first PoCs in 2014. After that operator’s participation has been spotty, suggesting that those heavily involved at the beginning of the process have moved on to private PoCs and trials. Since the Q1 2015, 50% of PoCs see direct operator involvement, ranging from Orchestration, NFVI or VIM with operators who are mostly new to NFV, suggesting a second wave of service providers are getting into the fray with a more hands-on approach.


Figure 36: Operators activity in PoC

Out of the 52 operators participating in the 38 Pocs, Telefonica, AT&T and BT, DT, NTT, Vodafone account for 62% of all PoCs, while other operators have only been involved in one PoC or are just starting. Telefonica has been the most active overall, but with all of its involvement in 2014, no new PoC participation in 2015. AT&T has been involved throughout 2014 and has only recently restarted a PoC in 2015. British Telecom has been the most regular since the start of the program with in average close to one PoC per quarter.



Figure 37: ETSI NFV PoC operators’ participation


On the vendors’ front, 87 vendors and academic institutions have participated to date to the PoCs, led by HP and Intel (found respectively in 8% of PoCs). The second tier of participants includes, in descending order, Brocade, Alcatel Lucent, Huawei, red hat and Cisco, who are represented in between 5 and 3% of the PoCs. Overwhelmingly, in 49% of the cases, vendors participated to only one PoC.

The most interesting statistics in my mind is showing that squarely half of the PoCs are using SDN for virtual networking or VIM and the same proportion (but not necessarily the same PoCs) have deployed a VNF orchestrator in some form.

Friday, September 4, 2015

Video is eating the internet: clouds, codecs and alliances

A couple of news should have caught your attention this week if you are interested the video streaming business.

Amazon Web Services confirmed yesterday the acquisition of Elemental. This is the outcome of a trend that I was highlighting in my SDN / NFV report and workshops for the last year with the creation of specialized clouds. Elemental's products are software based and the company was the first in professional video to offer cloud-based encoding on Amazon EC2 with a PaaS offering. Elemental has been building virtual private clouds on commercial clouds for their clients and was the first to coin the term "Software Defined Video". As Elemental joins AWS, Amazon will be one of the first commercial clouds to offer a global, turnkey, video encoding, workflow, packaging infrastructure in the cloud. Video processing requires specific profiles in a cloud environment and it is not surprising that companies who have cloud assets look at creating cloud slices or segregated virtual environment to manage the processing heavy, latency sensitive service.


The codec war has been on for a long time, and I had previously commented on it. In other news, we have seen Amazon again join Cisco, Google, Intel, Microsoft Mozilla and Netflix in the Alliance for Open Media. This organization's goal is to counter unreasonable claims made by H.265 / HEVC patent holders called HEVC advance who are trying to create a very vague and very expensive licensing agreement for the use of their patents. The group, composed of Dolby, GE, Mitsubishi Electric and Technicolor is trying to enforce a 0.5% fee on any revenue associated with the codec's use. The license fee would apply indiscriminately to all companies who encode, decode, transmit, display HEVC content. If H.265 was to be as successful as H.264, it would account in the future for over 90% of all video streaming traffic and that 0.5% tax would be presumably levied on any content provider, aggregator, APP, web site... HEVC advance could become the most profitable patent pool ever, with 0.5% of the revenues of Google, Facebook or Apple's video business. The group does not stop there and proposes a license fee on devices as well, from smartphones, to tablets, to TVs or anything that has a screen and a video player able to play H.265 videos... Back to the Alliance for Open Media who has decided to counter attack and vows to create a royalty-free next generation video codec. Between Cisco's Thor, Google's VPx and Mozilla Daala, this is a credible effort to counter HEVC advance.


The Streaming Video Alliance, created in 2014 to provide a forum for TV, cable, content owners and service providers to improve the internet video streaming experience welcomes Sky and Time Warner Cable to the group already composed of Alcatel-Lucent, Beamr, CableLabs, Cedexis, Charter Communications, Cisco, Comcast, Conviva, EPIX, Ericsson, FOX Networks, Intel, Irdeto, Korea Telecom, Level 3 Communications, Liberty Global, Limelight Networks, MLB Advanced Media, NeuLion, Nominum, PeerApp, Qwilt, Telecom Italia, Telstra, Ustream, Verizon, Wowza Media Systems and Yahoo!. What is remarkable, here is the variety of the group, where MSOs, vendors, service providers are looking at transparent caching architectures and video metadata handling outside of the standards, to counter specialized video delivery networks such as Apple's, Google's and Netflix'

All in all, video is poised to eat the internet and IBC, starting next week will no doubt bring a lot more exciting announcements. The common denominator, here is that all these companies have identified that encoding, managing, packaging, delivering video well will be a crucial differentiating factor in tomorrow's networks. Domination of only one element of the value chain (codec, network, device...) will guarantee great power in the ecosystem. Will the vertically integrated ecosystems such as Google and Apple yield as operators, distributor and content owners organize themselves? This and much more in my report on video monetization in 2015.