Reply to comment:

Michael Richardson
Thank you for this. I think that one of the key things to reducing e-waste is to keep devices in productive use longer. The smartphone cycle is largely being driven by planned obsolescence through lack of software updates. While few of us here are in the smartphone business, we see the same thing with CPE devices. No software updates means poor security, which pushes for replacement of the devices. More capable devices can sustain more updates, but cost more up-front, and there still very little relationship between price paid and number of years of supported software updates. At least, I believe this anecdotally. It would be nice if we had survey data to prove or disprove this belief. Perhaps RIPE along with some of the entities you mentioned might be able to do a confidential survey of ISPs in order to summarise the results? EN 303 645 and upcoming UK legislation requires that the support time for devices to be made clear at point of purchase. See for many references. (Yes, I'm the presenter for the webinar, the first of which was August 28) While the Turris system is well supported, and very capable, it does not seem have had a lot of influence on big ISP purchasing. What kind of software support lifespans are ISPs able to contract? While many RIPE clueful people know about, deploy and contribute to openwrt, it does not seem to show up in a supported way in devices that I see either big or small ISPs deploying. The big ones don't seem to care. The small ones can't afford to take a risk. Years ago, it was the common that an ISP would buy some 12-port switching device for it's core, and two years later (when it was too slow), would migrate it from the core to an access aggregator, and then two years later, it might find a few more years service as internal lab equipment, or being used for a multi-tenant CPE. I learnt this as a switch (chip) designer. Is this still a thing? In my more recent (2014-era) hands-on ISP experience, this was no longer a thing, because the devices had all become far too specialized. Maybe SDN is changing this? Are there best practices in making purchase decisions that support this migration of equipment? {I'm reminded of: }