Home / Blogs

Why SDN is Not Enough

Martin Geddes

A hot topic in telecoms at the moment is 'software-defined networking' (SDN). This term covers a range of technologies that put networks under the control of centralised management software. But what if SDN misses the point of why broadband networks exist in the first place?

Network equipment vendors are busy pushing operator CTOs to adopt a 'software telco' approach. A small army of analysts and consultants cheer this process on. There are legitimately claimed benefits: labour cost savings (through automation), and improved resource use (by dynamically allocating supply to demand). There are also serious technical risks [PDF].

The historical nature of the telecoms industry is one that is very focused on its internal technology mechanisms. Telcos like to sell megabytes, carried over megahertz. We talk to customers about bearer types (like fibre) and technology generations (like 4G). How the user puts the resulting data service to use is typically seen as a "not my problem" type of question.

The essence of SDN is to create a software model of the current data network business. This quantitative model is based on volumes of data: what 'bandwidth' resources do I have (i.e. supply), and how can I give different quantities of this 'bandwidth' to different users and uses (i.e. demand)?

This works well enough for core networks, where you have millions of "statistically aggregated" flows, so the quality has relatively low variability. In this case 'bandwidth' is a reasonable approximation to the customer experience being delivered.

However, nearer the network edge, a single average metric like 'bandwidth' fails to capture the rapid variation in packet loss and delay that affect the real customer experience. The performance of an application depends (among many other things) on a quantity of quality being delivered by the end-to-end network service. You can't just assume ever more 'bandwidth' is the answer to all your QoE problems.

If telcos want to be able to differentiate themselves from being commodity bandwidth 'pipes', they need to begin asking themselves some different questions:

  • What are my customers using my broadband network to achieve, in terms of business and application outcome?
  • What are the different levels of application performance that they are willing to pay for as a result?
  • What are the "quantities of quality" that enable that performance, and hence customer experience?
  • How can I create a network that delivers the right experience to each user and use (and at the right cost point based on willingness to pay)?

This requires a fresh approach. At the end of the day broadband networks only exist to deliver customer experiences, and not bandwidth. So rather than software-defined networking (SDN), I believe that we need a new software-defined customer experience (SDCX) approach.

Commercially, SDCX is a demand-pull model, rather than one based on supply-push. It makes relevant marketing promises that have value in business terms (like page load time). Technically, it requires new performance metrics that are strong proxies for the customer experience. Better metrics prove that that those marketing promises are being predictably delivered.

Crucially, a software-defined customer experience is not just "SDN with added metrics for jitter and packet loss". It is one that fundamentally puts the customer experience first, and the network then dynamically constructs that experience.

The problem with SDN is that it's automating a legacy supply-centric model based on circuits. It is (validly) solving an internal telco problem, not a customer one. The models don't relate to the customer experience closely enough, so SDN is limited to relatively modest goals. It cannot transform the role of telcos in the digital value chain.

The potential "big win" for telcos from SDCX is to become digital experience quality partners to enterprises. Rather than offering commodity inputs, they can offer a rich suite of capabilities to meet diverse cost and performance needs. As business partners, they can help to surface trade-offs of cost and customer experience in the "digital transformation" planning process.

What is perhaps most surprising about SDCX is that the idea of putting the customer experience ahead of network technology is seen as radical and new. This tells you a lot about the maturity of the broadband industry, and how far it has yet to progress. The biggest opportunities for transforming society through data networks may still be ahead of us.

By Martin Geddes, Founder, Martin Geddes Consulting Ltd – He provides consulting, training and innovation services to telcos, equipment vendors, cloud services providers and industry bodies. For the latest fresh thinking on telecommunications, sign up for the free Geddes newsletterVisit Page
Follow CircleID on
Related topics: Networks, Telecom
SHARE THIS POST

If you are pressed for time ...

... this is for you. More and more professionals are choosing to publish critical posts on CircleID from all corners of the Internet industry. If you find it hard to keep up daily, consider subscribing to our weekly digest. We will provide you a convenient summary report once a week sent directly to your inbox. It's a quick and easy read.

I make a point of reading CircleID. There is no getting around the utility of knowing what thoughtful people are thinking and saying about our industry.

Vinton Cerf, Co-designer of the TCP/IP Protocols & the Architecture of the Internet

Share your comments

SDN is a solid start Peter Fretty  –  Feb 09, 2016 7:24 PM PDT

I agree SDN is only the beginning. There is a true need for an infrastructure-wide evolution. Bottom line as organizations head into the digital economy and then the IoT environment there needs to be a higher level of seamlessness in how we leverage technologies. Embracing a unified group of vendors could help as well — something that surfaced as a necessity in a recent IDG survey, especailly when considering security. Peter Fretty, IDG blogger for VMware.

Demand drives all Michael Elling  –  Feb 16, 2016 11:15 AM PDT

Martin, take it a step farther.  It is a bet on perceived demand (ex ante) in competitive markets that drives the investment (capex & opex) process, or supply.  But this cannot be done with a partial view of demand; be it one sided (originating or terminating) or geographically or contextually constrained.  That's the problem with core (OTT) vs edge (ISP) today.  You are beginning to come over to my way of thinking. 

After realizing this, then we need to establish settlements which serve as price signals to provide incentives and disincentives be they north-south (app to infrastructure) or east-west (between actors/users/networks/publishers, etc...). 

The current winner takes all model of balkanized actors is not sustainable.

Hopefully we get together when you visit NYC area again.  Best, Michael

To post comments, please login or create an account.

Related

Topics

Domain Names

Sponsored byVerisign

IP Addressing

Sponsored byAvenue4 LLC

New TLDs

Sponsored byAfilias

Cybersecurity

Sponsored byVerisign

DNS Security

Sponsored byAfilias