[ANNCMNT] Major LinkedIn Updates (that impact you!), Group Transition, Edge Computing, SD-WAN Standards, 4G vs 5G, Network Transformation Myths and IT Realities!

Hello Etherneteers!

Having given you an idea of the LinkedIn changes that affect both you and the Group; see here if you missed them http://bit.ly/2LC1cIs, it’s now time to jump into technology matters!

In keeping with tradition, in this update I am going to provide you a commentary on the discussions that spanned the Group between about Sep. 23- Oct 14. My next update will cover the period Oct. 15 – Nov. 14, and the one after that will cover Nov. 14 – Dec. 15, and that will bring us up to speed!

Why this Summary and Mindmap? And How to Use Them …

These summaries and the accompanying Mindmap http://bit.ly/2sIeUCY (Note: For some reason, you have to refresh your browser link at least once after clicking on the link above, for Adobe Acrobat to kick-in and for you to see the mindmap, so don’t forget to do that – you also need Adobe Reader and Flash Player installed to be able to use the Mindmap properly.) serve, in my view, a very useful purpose, and I’d like to take some time to highlight this, so folks can make maximum use of them. 

So, here goes:

  • First, these provide a handy compendium of significant topics discussed in the Group over a period of time. This gives you a wealth of knowledge at your fingertips, grouped logically by subject, with summaries and commentary. So, a well-curated and timely map that you don’t have to make the effort to search out or compile.

 Just imagine the amount of time and energy it would take for every  single one of you to do this completely on your own – you would first have to decide on a list of topics to study/research, then you’d have to identify reliable sources of information on these topics, then compile a reasonable amount of information on the topics from potentially multiple sources, then distill that information down to a few articles/writings/papers on each topic, and only then could you even being to read-up on a topic and start learning about it!

All this work is now eliminated. Rather, you can use this summary and Mindmap as a springboard to learn more about the areas that truly interest you.

Note that the Mindmap is clickable http://bit.ly/2sIeUCY, and includes “Notes” associated with the topics, which summarize key observations from that post/topic. Also included, are additional resources, such as pointers to articles, videos, whitepapers and such – useful to learn more.

  • Second, you can use this Mindmap and this blog post to navigate a large number of topics easily and effortlessly, focusing on those that interest you, or are relevant to your job/work and useful for expanding your technical skills. 
  • Third, due to the excellent quality of material shared on our Group (thanks to all of you!), simply browsing, even briefly,  the topics covered and their summaries, will substantially increase your awareness of technical and business developments in our industry, and keep you on the cutting edge.
  • Fourth, I’ve broken the material down into labeled sections that give you a hint of what’s they are all about – so you can easily skip to the ones of interest to you, without loss of fidelity. While the post may be long, my goal is that every member be able to extract value from it easily. (In fact, if you have thoughts on a better presentation, do share them in the comments here or on the Group in comments under the accompanying [ANNCMNT] post.)
  • Fifth, for as long as we are on the LinkedIn platform, this summary and Mindmap save you from the hassle of continuously scrolling down the Group’s main page to keep looking for articles/posts of relevance.  Without a good search feature available and with no ability to bookmark, LinkedIn has made it virtually impossible for Group members in active Groups like ours to easily find or track posts. 

This post and the accompanying Mindmap give you “one-click” access to the most useful discussions that occurred in the Group any given period. I’m not sure about you, but I find that of immense value – not only can I see related articles in context, I can also see their inter-relationships, jump to the articles and any associated resources in just one click, and never have to flounder or scroll endlessly looking for articles of interest! Or, looking for an article/post you saw briefly at some point (but didn’t have time to read immediately), noted that you would come back to it, but could never find it again!

To me, that’s worth its weight in gold! (Which, BTW, is saying a lot now-a-days, if you’re following how the fiat currencies worldwide are rapidly eroding in value due to inflation and low (or negative in some cases!) interest rates, leaving very few true stores of value.)

Note that the structure of the post below follows the structure of the Mindmap, so the different sections of the Mindmap are discussed chronologically in the post, starting with the very first section (and its subsections) and moving on sequentially to subsequent sections and sub-sections.


1. Thanks to Key Contributors

Folks, as you know, my motto is “it is the active participation of the members that makes the Group!”, so I always start by acknowledging the many members, who by their active sharing of knowledge, questions, and thoughts keep the Group energized, and make it a venue for assimilating so many insights about our industry.

As always, a big “Thank You!” to all of the contributors for this period for your participation, engagement, and contribution!

Posts by:

Gary Kim, Tom Nolle, Ahmad Tajuddin Samsudin, Michael Spence, Anton Karneliuk, Christopher Cullan, David Strauss, Victor Ma/Daniel Bar-Lev, Patrick Lopez, Michael Lemm, Vishal Sharma, Pav Solanki, Robert Alcock

Comments by:

Darell Merschak, Vishal Sharma, Ahmad Tajuddin Samsudin, Tom Nolle, Kevin Vachon, Christoph Jaggi, Anthony Senter, Glenn Kimble

Likes by:

Greg Phillips, Himanshu Chandra Tyagi, Jagdish Kumar Kaushik, Darell Merschak, Marjory Sy, Ian Horowitz, James Bensley, Hemant Bagga, Gunter Van de Velde, Nikolay Abromov, Modestas Rope, Anton Degtyarev, Muhammed Sozer, Raul Flamenco, Dedy Purnairawan, Javier Ruiz Gomez, Raghuram Srinivasan, Sheila Harrell, Harindha Fernando, Olga Aparicio, Andre Luiz, Solcia, Serkan Kaya, Edwin Quinones Jr., Alexandre Prandini, Nilam Bhate, Johan Hammarstrom, Frank Muche, Dipesh Patel, Pushkar Pal, Mark Manning, Michael Chapman, Mr. Farhan Mohamed, Michale Knoll, John Shepler, Vishal Sharma

And, now on to the round-up!


2. 5G & Mobile Evolution

5G Speeds …

5G access being heterogeneous will use a mix of frequency bands and aggregated frequencies. This multitude of ways of supplying 5G access means that performance will vary depending on the type of access, which will make comparing “5G” speeds challenging.  Non-Stand Alone (NSA) 5G (which means 5G that also uses 4G technology) will be different than standalone 5G. http://bit.ly/341hEZo

Also, it appears http://bit.ly/36hq6Ff that the  peak data rate of 5G at about 20 Gb/s is about an order of magnitude more than that of 4G at 1 Gb/s, and it’s latency at 1-10 ms can be about an order of magnitude less than that of 4G at 10-30 ms.  Keep in mind, these are all computed/tested assuming ideal conditions; real users in all three network types; 3G, 4G, and 5G, would not, by a long shot, achieve these ideal rates. This evolution, per Gary Kim, is in line with how the earlier generations of mobile technology have improved over the preceding generations. 

There are a few questions I had on the chart in Gary’s post, which I’ve posted to the Group so check those out there.

(Also note that there can be conditions under which 4G LTE outperforms the so-called “5G”, and that “5G” capable phones today are in the range of US $1000.)


… 5G Technology – User Plane & Network

A very apt observation by Tom Nolle http://bit.ly/2PuEVxx was that  when considering 5G UPF (user plane function) virtualization, it is preferable to think of an abstract black box implementation instead of decomposing the contents of the black box (each of the individual  UPF elements) and making those individually visible.  

This is akin to the two approaches that evolved for 4G LTE EPC virtualization – virtualization of the whole EPC (so that it presents to the external world standard interfaces, but could have different realizations inside) versus using VNFs to replace the individual elements (say, the SGW, PGW, MME, and eNodeB), which provides less flexibility.

This http://bit.ly/2PnyspI is an excellent article discussing what changed about the network’s mission with the advent of data, and discusses the change in architecture this motivates. What would eventually emerge is an architecture that consists of an “access zone” and a “service zone” with traffic aggregating towards the “service zone”. This would lead to a mesh of low-layer (optical or electro-optical) connections between large access routers, with traffic being sent edge-to-edge, and the role of core routers being progressively deemphasized.


Business Implications for Fixed Wireless and Wi-Fi Offload

Per a Parks Associates survey http://bit.ly/2omp7TQ that Gary Kim discusses, which it should be noted did not take cost into account, users think faster 5G  fixed-wireless service would be an attractive substitute for their existing fixed-line (but slower speed) Internet access. 

The key observation is that unlike 5G mobile phone service, 5G fixed wireless does not require investing in new handsets, especially if the ISPs make customer premises gear available at low recurring charges or at subsidized prices at terms familiar to customers.

An interesting question is http://bit.ly/32TkOyo: if with 5G, mobile access speeds and realized latency are equal to or even better than that of fixed access, and, if the tariff structure remains the same as it is today, would there still be a role for Wi-Fi offload?

This is because in many markets the users may shift to using their mobile connections directly (as opposed to Wi-Fi). So would operators still want customers to use Wi-Fi where possible?

The answer is “Yes”, especially for MVNO’s that have to buy capacity wholesale, and, therefore, would be sensitive to how much of that gets used! 


5G Revenue Musings

So how much of the total mobile operator revenue will 5G generate? Gary Kim, summarizing a report by Strategy Analytics http://bit.ly/2Nk3uw6, observed that even if half of all phones sold by 2025 are 5G-capable, 5G will still generate only about a quarter (25%) of all wireless service revenue. 

The key is the time-frames in which low-priced and mid-range handsets will be available. The faster this happens, the faster 5G adoption by consumers is likely to be.

Finally, I find it very surprising that even though we don’t have any visibility into the impact of 5G on service provider revenues http://bit.ly/31UdFfV, 5G is being pushed so heavily in our industry at present, with nearly everyone and their brother atop the 5G bandwagon.

The popular sentiment is that revenue upside might come from higher-value services for businesses incorporating: the Internet-of-Things, edge computing, virtual private networks, artificial or augmented reality, but it is not clear that these will take off to a degree where ARPU’s are seriously boosted any time soon.

How can operators sink billions into 5G (when they haven’t even amortized their 4G investments, and, in some cases, 3G investments yet!) with no visibility into and no plan for how 5G will shore up their revenues and ARPU?


3. “Cloud Native”: What is that?

Evolution …

What constitutes “cloud-native”, and what can it do for the industry? To find out check out Tom’s article http://bit.ly/2PqYDKt, where he explains that “cloud-native” has the potential of adding $1 Trillion a year to overall IT spending, and increase vendor revenues in the hosting, software, and virtual networking spaces by up to 50%.

VMWare and IBM are the big players competing in this space, but HPE and Microsoft may soon join in, and all of the vendors are focusing on building eco-systems of cloud-native products, which would eventually give the enterprise customers the products and support they need.

How do we know which applications justify cloud-native treatment? For that one has to know both the benefit that the application will bring, as well as the benefit that cloud-native will bring to the application http://bit.ly/2YrGVuy.  

You’re best served by reading the entire article, but here’s a hint: Cloud-native technology works best when applied to “events”.  A transaction can be dissected into a series of “events”, each representing a step a user takes in interacting with a commercial application. The processes for each step can then be made so simple as to be stateless and fully scalable and distributable, which allows such processes to be put in the cloud using techniques like VMs, containers, micro-services or combinations thereof.

Despite the hype around the “cloud’, Tom Nolle in his post http://bit.ly/38gp8uV.  observed that the cloud-native model isn’t well understood anywhere – not within enterprises and even less so within operators! Operators are worried about issues like:  ballooning opex with virtualization, VNF onboarding, the management of virtual functions, and the reliability and availability of cloud-native deployments. So they are looking at partnerships with cloud providers to help realize their own near-term opportunities for cloud-hosting of features and functions.


… And Business Cases

It turns out that enterprises and service providers are both having difficulty meeting their corporate ROI expectations http://bit.ly/2YwBieI. For operators, it is the decade-long issue of increasing the profit-per-bit; for enterprises, it is the notion of making IT investments to lower cost, which is running out of options, because high-ROI IT projects are bitten off first, leaving low-ROI projects dangling.  Some of the reasons for this phenomenon, and possible solutions are outlined in the Mindmap notes, and covered in the post. Hint: Cisco notes that 5G isn’t generating new opportunity in the near-term. So, for the ROI expectations to be addressed, buyers must take more control, perhaps via open-source projects and more “cloud-native” deployments.

On the same note, enterprise CIO’s are still (in 2019) figuring out how to make a cloud business case http://bit.ly/2rtR0eb. Until a couple of years ago, most just accepted the touted benefits of the cloud – the lowering of “cost”, the consolidation of “compute”, the “operations economies of scale” supposedly resulting from managing applications in the cloud, and so on.  

As per Tom Nolle, however, the capex+opex savings from application migration to the cloud would make sense for less than 10% of applications traditionally run in an enterprise data center. But with appropriate adaptation for a “cloud-native” environment, another 65% of enterprise applications could be migrated profitably to the cloud. To truly benefit from the cloud, therefore, enterprises would have to deploy applications in a way that improved productivity and sales, which is outlined in detail in Tom’s article.


4. SDN/NFV: Still a Conundrum?

Simple? Or Complex? …

Our member, Ahmad Tajuddin Samsudin, from Telekom Malaysia raised highly pertinent issues on the complexity of SDN and NFV deployments in a service provider environment. http://bit.ly/2PqYDKt Ahmad argues that SDN and NFV may reduce opex, but not capex. Further, there are 4 factors for an operator to consider before choosing an SDN/NFV solution: interoperability, software licensing cost, operational skills/competency needed, and technologies available. Check out Ahmad’s commentary here http://bit.ly/2PqYDKt.


… and Thoughts on Business Cases: TCO and Per-VNF Savings

Is “virtualization” a cost-savings strategy for operators? It’s hard to know, given there are many different views of what “virtualization” is! For this, one needs to see which camps operators fall into, which network segments different popular virtualization strategies (SDN, white-boxes, etc.) fit into (access, transport, data center), and the virtualization architecture that they adopt. All of these are detailed here  http://bit.ly/2Pr8nWF.

There are 3 reasons why NFV fails to reduce opex and deliver the savings it was touted to provide, and this is a key problem with NFV’s business case, as detailed here http://bit.ly/2qV2OG1.  Operational complexity of VNF onboarding and integration, licensing cost of VNFs (which Ahmad also observed http://bit.ly/2Pr8nWF), and non-realization of the expected economies of scale. To see how these three problems can be fixed, see this post http://bit.ly/2qV2OG1.


5. Enterprise Architecture … is Evolving!

Michael Spence has a useful post that discusses how current enterprise architecture is being shaped by SD-WAN http://bit.ly/34428fh. He compares enterprise connectivity as provided by MPLS or modern SD-WAN solutions, and lays out the criteria that must be satisfied at the branch office level for enterprises to digitally transform themselves.

Amazon’s Sidewalk protocol for IoT devices that lets people build IoT networks using relatively cheap devices over fairly large ranges (of about half-mile) is the subject of this post http://bit.ly/2MV0NCg. Sidewalk joins several other existing protocols for such communication: Zigbee, Bluetooth, WiFi, NB-IoT, and Insteon.

The goal of IoT protocols is to allow cheap sensors to connect to controllers over a network, and Sidewalk focuses on cheap, simple, and secure connectivity, and integrates with devices in Amazon’s Ring family/system. The broader goal for Amazon, it appears, is to eventually link multiple facility networks into a federated complex.


6. Network Automation: A Panacea?

Understanding it …

Network automation is in the air now-a-days, and we have a number of posts on the topic by Anton Karneliuk, who discusses what network automation is http://bit.ly/32V0U61. In fact, new member Kalyan Kidambi, who joined recently, has spent considerable time looking at solutions for network automation, and may have some inputs on Anton’s posts ☺.

David Strauss wrote a post for the MEF http://bit.ly/36hJaDq, where he discusses a survey of cable MSO executives on their network automation strategies.  The key findings are discussed in David’s article, but here are some highlights: operators’ initiate network automation to improve efficiencies and lower opex because it is believed to bring the broadest benefit to the company, and it touches the most parts of the company – systems, groups. 

The other key driver for Tier 1’s is to reduce human errors, while for Tier 2’s it is to improve customer experience.  Some of the gaps that operators pointed to were lack of interoperability, the software not being advanced enough, and decision makers still not being convinced of its value.


… and Experimenting with It

Anton also shared a number of posts on the Group dealing with different aspects of network automation, which he demonstrates live in his videos. See for instance, segment routing in the data center http://bit.ly/31Qqrw9, closed-loop network automation http://bit.ly/36gbGVN, and  whitebox switches in the data center http://bit.ly/2PuLBMd.


7. Networking: Real or Virtual?

A software-hosted router network could be realized either as a network of software-hosted routers (each of which, from the outside, looks exactly like a real router would look) or as a software-implemented router network (which, from the outside, would look like a router network, where each individual router wouldn’t be visible as an abstract model). NFV’s original goal in 2012 was to reduce capex by substituting proprietary devices with software instances hosted on commercial devices, but soon lead to the realization that this idea alone would not cut capex significantly. Thus, a better notion of virtual routing is needed, as discussed in this insightful post http://bit.ly/2LAEyQR.

Despite virtual networking being a foundation for initiatives like NFV, it is still not clear what a virtual network is http://bit.ly/2E1f2jz? What is clear is that a network of virtual devices is not the same as a virtual network. To enable us to understand the characteristics of a virtual network, the article proposes four principles, which are worth understanding. http://bit.ly/2LAEyQR.


8. Telco Evolution: Where are we headed?

Telco Planning …and the API Economy

The MEF’s LSO Sonata SDK, released in June 2019, establishes a framework for inter-provider service automation of all MEF 3.0 service offerings, including IP and SD-WAN services.

An important component of business-to-business automation includes the ability to assess the capability of potential partners to deliver a network service, based on their reputation and past history—quickly and automatically. Technologies like blockchain, coupled with LSO standard APIs enable this to become a reality. 

In this post http://bit.ly/2PteZ5g Daniel Bar-Lev of the MEF and Victor Ma of DCConnect discuss the use of blockchain and smart contracts to establish a type of inter-carrier credit rating that carriers can use when evaluating which partner network to use to route services over. 

The idea is to offer wholesale operators the ability to build a proven reputation that distinguishes them from their competitors. For example, measuring the very important parameter—How long does it take the operator to deliver their wholesale Carrier Ethernet service once it is ordered via LSO Sonata? 

To learn more check out this interesting discussion here http://bit.ly/2PteZ5g, and do share your thoughts and questions.


… with a Coherent Value-Chain

Meanwhile Patrick Lopez http://bit.ly/2q5LZYb posits that the main components of edge-computing are: compute, storage, connectivity, and continuity – the notion that the edge of a cloud or a network is the closest point to the user / developer to provide a continuous and consistent experience. 

Patrick explains some interesting ideas behind what it takes to provide such a continuous experience and how the telcos/operators can partner with cloud providers to use edge-computing to re-enter the cloud value chain, by providing valuable performance enhancement based on an asset that is expensive and difficult to duplicate – their network infrastructure. So check that out here http://bit.ly/2q5LZYb.

And, if 5G’s promise to deliver a substantially different user experience (not just more speed / volume, but different services with guaranteed performance) is to bear fruit, edge computing will be an important part of the deployment options.


… SD-WAN Prospects: Business Cases and Observations

Michael Lemm shared http://bit.ly/2MWVguW  an article penned by Sorell Slaymaker , which put a stake in the ground stating that SD-WAN can reduce enterprise WAN costs by half.  This view, however, was deemed limiting by several Group’s members who commented that SD-WAN has advantages beyond just saving costs.

First, it is not clear that just saving cost is the key reason enterprises migrate to SD-WAN. As Ralph Santitoro noted, there are other considerations, such as the prospect of getting high-availability, significantly more bandwidth (via separate underlay connectivity services), and, by using different optimization techniques (packet reordering, data compression/caching/de-duplication, TCP proxying, and optimal path selection based on near-real-time measurements of UCS QoS performance metrics) , better application performance and user QoE.

In the Mindmap, I have provided pointers to a few excellent presentations by Ralph on the MEF’s SD-WAN service and the recently released MEF 70 document.

Tom Nolle had a different view, stating that in a majority of the cases SD-WAN can save money versus MPLS-VPNs, if used in the “replacement model”. It turns out, however, that in many cases SD-WAN is used to supplement MPLS VPNs in locations where they are not available or not affordable. 

Also, MEF 70, in my view, harmonizes terminology, and provides some characteristics of an SD-WAN service but cannot really be used as a definition of SD-WAN itself.

Finally, to understand how SD-WAN dynamics at both the enterprise and the operator have changed in just 2019 alone, check out this post http://bit.ly/2P4Mv2E.  There is wealth of insights in there! 

Some highlights: while the high-cost of MPLS (per 60% of enterprises polled) and unavailability of MPLS at all sites (per 70% of enterprises polled) is a reason to go to SD-WAN, for most companies SD-WAN is an adjunct to MPLS VPNs. So, very few are completely swapping out MPLS VPNs to go with SD-WANs.

On the operator side, the primary driver for SD-WAN offerings is competition (for over 75% of operators polled), gaining additional revenue from sites unreachable via MPLS (66% of operators polled), and to contain the impact of connectivity problems created by the combination of MPLS and SD-WAN (33% of operators polled). The key for operators is to raise SD-WAN revenues and attempt to limit the risk of MPLS VPN displacement. Some interesting dynamics are discussed in the article http://bit.ly/2P4Mv2E.


9. Industry Goings-On!

SDN/NFV World Congress 2019 … Where it all began back in 2012!

The SDN/NFV World Congress took place in The Hague, Netherlands in October 2019. Pav Solanki posted about the NTA Award nominations here http://bit.ly/2rvQlZV (I had the privilege of being one of the Judges –  thoroughly enjoyed the exercise). The Mindmap has a link to all the video recordings of the conference, so do check those out.

Thanks Pav Solanki, Millad Farahani, and the rest of the Layer123 team for putting together some great events!


… to the Crossover between Telecoms and Broadcast …

Robert Alcock shared an interesting video http://bit.ly/2NxXwb9 of a panel discussing the advances in the broadcast industry due to its use of IP (Internet Protocol) networks for broadcast, and that industry’s adoption of software. Thus, a great confluence of telecoms and broadcast, which is changing the face of the broadcast industry and how QoE is measured – bringing in both qualitative and quantitative measures. 

The Mindmap has a detailed summary that you can check out, and, of course, watch the video.


 … to VoIP and the Independent Telco!

Finally, our member Andrew Ward organized a virtual conference on “VoIP and the Independent Telco” http://bit.ly/2WoiJZ6;  I had the pleasure of participating in the session “TDM, VoIP, Virtualization and the Future of the PSTN”. 

We tackled questions such as: what is the role of the independent telco in current times? How relevant is it given the prevalence of OTT services? What is the role of virtualization for small telcos? How do they transform themselves to benefit from current trends? Where is revenue going? And what have some existing telcos done to beat the trends … so quite a range. 

The actual conference tackled substantially more issues, such as voice network architecture, security and toll fraud, Perimeta training, and making a hosted PBX product work better, all of whose details you can find here http://bit.ly/2nWOWJP

For those that missed the live event, you’re in luck! Recordings of the entire conference’s sessions (5 days) are available for just $99 for an individual purchase, and $250 for a corporate purchase, through http://bit.ly/2nWOWJP, just scroll to the bottom of the page for the order form.

Once again the Mindmap corresponding to this update can be found at http://bit.ly/2sIeUCY and remember to refresh your browser page at least once, if the Mindmap does not show up the first time! (Yes, quirks of Adobe, Flash, and html, I know .)

Until next time … may the bits in your bytes and the bytes in your packets be profitable! ☺

Any questions or observations?  Let me know! Remember, I’m standing for all that is possible for you in telecom networking and professionally.