Talk:Software-defined perimeter

Needs wikification

It seems that some of the text was copied and pasted from somewhere, e.g. leaving references like "[13]" and including lots of non-wikified references. At a minimum a lot more wiki links should be added. ★NealMcB★ (talk) 15:44, 17 May 2015 (UTC)

Too much promotion via poorly-explained features and unsupported claims

This reads like an advertisement for a concept. It needs lots more information on the networking and protocol techniques used, and reviews and evaluations from third parties about this technology. ★NealMcB★ (talk) 15:53, 17 May 2015 (UTC)

black ip addresses

The article claims: Application infrastructure is effectively “black” (a DoD term meaning the infrastructure cannot be detected), without visible DNS information or IP addresses. But based on the spec, surely the controller IP address is typically visible via DNS, and the ip addresses for servers could be discovered from network traffic). ★NealMcB★ (talk) 16:33, 17 May 2015 (UTC)

Agree, it is black in the sense of being ciphertext (covered or encrypted data, equivalent to unclassified), vs plaintext which is red. Black does not mean 'invisible' IP/DNS. — Preceding unsigned comment added by 141.168.237.73 (talk) 17:04, 6 February 2019 (UTC)

Actually, an invisible IP regime and DNS entries are not all that farfetched. SPD looks to me essentially like network micro segmentation down to the device level, with a couple of novel concepts thrown in. Several vendors are already offering network authentication systems and policy enforcement at the port level. This essentially gets you invisible IP address regimes, as the device can only talk with the layer three switch port to which it is physically attached and sees nothing else in the network until authentication and authorization are complete. Considering the few novel concepts, a number of security and network companies are offering micro-segmentation gateways that sound suspiciously like the SPD gateway host described in the article. SPD controllers sound similar to offerings for network level authentication systems already on the market. The only other novel concept is the SPD enforcement agent that apparently every server, client and piece of infrastructure has to have loaded. Once one grants that all servers will have an SPD agent, essentially a distributed authorization aware firewall, is it so far fetched that one could integrate such an agent into DNS and thereby only answer requests which the device/user is authorized to see? --LordGeep (talk) 04:48, 30 August 2019 (UTC)