A Tier-1 deployed it today. http://www.lightreading.com/data-center/data-center-infrastructure/apstra-de... It can be a lot of things and nothing, it can be just a GUI with some UX built into it, relatively making it easy to configure stuff, and it can be just a bunch of well written automation scripts with all sorts of checks and whistles, it can be just a little bit more analytics on top of that, and it can definitely be something a lot more if done PROPERLY, I hear good things about Apstra but i cant confirm because it can easily be marketing. In my humble opinion, if done properly, the Network Operating System would send telemetry info on any change in the network, and the Intent Based Networking system would consume that, run it by a lot of policies and take automated/guided actions. There is a webinar tomorrow with their VP of Engneering, join it and ask your question: http://go.apstra.com/webinar-apstra-cumulus-disaggregation Cheers, Marcus
On Tue, May 29, 2018 at 7:29 PM, K. Scott Helms <kscott.helms@gmail.com> wrote:
A substantial percentage, perhaps 100%, of the use case can be accomplished using micro-segmentation.
On Tue, May 29, 2018 at 2:33 PM LF OD <bz_siege_01@hotmail.com> wrote:
Been following the articles on "intent-based" networking from Cisco and other vendors for a couple years now, and I have a basic grasp of the concept of "define your goals/outcomes and automation will make it so", but I do not know the practical applications of it or how you are supposed to convey your "intent". However, $dayjob is a medium-sized enterprise and looking at a potential refresh in the upcoming budget.
So... do any of you at ISPs, Cloud Providers, or Enterprises have any hands-on experience with "intent-based" networking products? If so, could you please provide some info on what caused you to look at it, who did you look at, why did you choose whoever you chose, and is it working out for you?
If you looked at it very closely and decided not to bother with it, I'd also be interested on your take. At $dayjob we have been adopting automation at the application and workload level, but there hasn't been much need for it at the infrastructure level. However, if we can improve the services we offer while refreshing that infrastructure, it wouldn't hurt - depending on cost of course.
Thanks in advance.
LF0D