My opinion is that the current stance of ‘deny all’ is probably the safest bet for all parties (including users) at this point. It’s been that way for years now, and is a substantial change that may result in little benefit. After all, you’re probably looking at most users removing the default rule(s) just to add something that’s more restrictive and suits their organization’s security posture. If they aren’t, then it’s possible they’re introducing unnecessary risk.
There should be some onus put on the provider and/or the user/project/tenant to develop a default security policy that meets their needs, even going so far as to make the configuration of their default security group the first thing they do once the project is created. Maybe some changes to the workflow in Horizon could help mitigate some issues users are experiencing with limited access to instances by allowing them to apply some rules at the time of instance creation rather than associating groups consisting of unknown rules. Or allowing changes to the default security group rules of a project when that project is created. There are some ways to enable providers/users to help themselves rather than a blanket default change across all environments. If I’m a user utilizing multiple OpenStack providers, I’m probably bringing my own security groups and rules with me anyway and am not relying on any provider defaults. James On 3/2/16, 3:47 PM, "Jeremy Stanley" <fu...@yuggoth.org> wrote: >On 2016-03-02 21:25:25 +0000 (+0000), Sean M. Collins wrote: >> Jeremy Stanley wrote: >> > On 2016-03-03 07:49:03 +1300 (+1300), Xav Paice wrote: >> > [...] >> > > In my mind, the default security group is there so that as people >> > > are developing their security policy they can at least start with >> > > a default that offers a small amount of protection. >> > >> > Well, not a small amount of protection. The instances boot >> > completely unreachable from the global Internet, so this is pretty >> > significant protection if you consider the most secure system is one >> > which isn't connected to anything. >> >> This is only if you are booting on a v4 network, which has NAT enabled. >> Many public providers, the network you attach to is publicly routed, and >> with the move to IPv6 - this will become more common. Remember, NAT is >> not a security device. > >I agree that address translation is a blight on the Internet, useful >in some specific circumstances (such as virtual address load >balancing) but otherwise an ugly workaround for dealing with address >exhaustion and connecting conflicting address assignments. I'll be >thrilled when its use trails off to the point that newcomers cease >thinking that's what connectivity with the Internet is supposed to >be like. > >What I was referring to in my last message was the default security >group policy, which blocks all ingress traffic. My point was that >dropping all inbound connections, while a pretty secure >configuration, is unlikely to be the desired configuration for >_most_ servers. The question is whether there's enough overlap in >different desired filtering policies to come up with a better >default than one everybody has to change because it's useful for >basically nobody, or whether we can come up with easier solutions >for picking between a canned set of default behaviors (per Monty's >suggestion) which users can expect to find in every OpenStack >environment and which provide consistent behaviors across all of >them. >-- >Jeremy Stanley > >__________________________________________________________________________ >OpenStack Development Mailing List (not for usage questions) >Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe >http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev