The Post-SDN World
What's the Next Big Thing?
It is way too early to start seriously planning for the post-SDN telecomm world. Real SDN solutions are just starting to take off and there are still many gaps in reaching the full potential of SDN. However, if you have been in the industry for a long enough time, you know that all of these trends go in cycles so there will come a day with SDN is commonplace (or abandoned) and everyone will be excited about the next big thing. I’ve been asking service providers and other smart minds in the industry for a few weeks now what they think will be the next big thing in telecomm.
The best answers that I’ve gotten all seem to fall into the general bucket of big data analytics. Once we get to a more open, programmable, automated, and intelligent network, we will have access to more information than ever about not just the current performance of the network, but also how the network is being used and paid for. Network operators will be able to view historical information on network usage down to the minute if they want to, and will be able to understand in minute detail exactly what their customers are willing to pay for and when. This will lead to an explosion of information that needs to be sliced and diced so that it can be poured back into the network to provide better solutions for everyone.
Analytics are already a big part of the SDN evolution. For example, advanced path computation engines that take into account network usage, operator routing preferences, customer priorities, and more. Analytic based load balancing engines which claim to be able to re-balance networks based on performance measurements. Gathering data from more intelligent network elements can be used for predictive maintenance and advanced alarm/degradation correlation to improve network uptime. So we are already moving in the direction of big data.
However, as Yaakov Stein, CTO of RAD, explained at this year’s MEF17 conference, the information that we are gathering may not actually relate to the user experience. In the future, advanced data gathering techniques will give us even more information about the user experience that service providers will need to parse and analyze.
A truly automated, fully programmable, application-aware network environment can offer orders of magnitude more information than current networks. Anyone who spends time with data analysts knows that the biggest problem with big data is determining what part of the data is useful out of the reams of noise. A solution that, for example, can scan through the noise to determine that customers are willing to pay more for solution Y than for solution X under specific conditions could be invaluable. Likewise, a solution that could determine that deploying technology A in a particular location could reliably avoid network downtime or ensure additional customer acquisition would also have a major impact.
These days, the good news is that data storage is relatively cheap so almost anything that can be measured can be stored. The bad news is that data storage is relatively cheap so almost anything that can be measured can be stored. As we get more intelligent networks, we will get more data. The post-SDN world is starting to look like a world run by very smart data analysts or an artificially intelligent network which can measure, gather, store and analyze the data.