Is there a way for teams with production Docker deployments to easily experiment with this kind of scanning on their own infra to understand their own situation? Maybe worth writing up a quick description of how operators can do something like that.
We are using Grafana with OpenTSDB and loving it -- really looking forward to some of the 2.0 features and about to buy a bunch more TVs for big shiny Grafana dashboards in our NOC. Way to go guys!
they are essentially the same thing -- both are "proprietary" names for the same feature, which is behind-the-scenes recursive CNAME chain lookups by the authoritative nameserver, to return A records directly.
you're correct about ALIAS (although practically, it doesn't matter: people are going to use the apex whether it's proper or not at this point). i'm more referring to other complex usually-proprietary capabilities of big dns providers, especially traffic routing features. routing semantics are generally not translatable across providers, and if you're using dns based routing (as most cdns, major web properties, etc are) then doing multi-network dns gets a lot harder. if you're amazon, you write and maintain a bunch of code to span providers. if you're not, the barrier to multi-network is high if you're doing more than static dns.
http://blog.cloudflare.com/introducing-cname-flattening-rfc-... is a reasonable explanation. fundamentally a CNAME says "when you get queries for this name, go look at this other name instead". among other things, doing a CNAME at the zone apex means resolvers can't then find your NS, MX, or other records at the apex, which is problematic.
good luck finding any major online property or infrastructure that isn't making use of some kind of proprietary syntactical dns sugar. it doesn't mean you can't span providers, but it does mean it takes a lot more work to do so.
anyway, you're not wrong, the best approach to mitigate this kind of thing is to leverage multiple dns networks. but doing so is not easy unless the application is still using dns like it was in 1995, and that is increasingly rarely the case.
hi. i did not mean to spam or kick dnsimple, we know them and they are a great company and service. we are actively receiving inbound queries about this from folks asking for help, so thought it made sense to chime in publicly here. but you're right, i should have kept it on topic to the discussion at hand instead of offering anything else up.
you're not wrong: in this industry you never kick your competitors when they're down, everyone is subject to the same constraints, attacks, and complications. that wasn't my intention and i said so in the post.
dns is less easily distributed when fancy features like ALIAS (which dnsimple is widely known for) are in the mix. and wide distribution isn't enough to win vs truly volumetric attacks. it takes a lot of ports and compute to absorb 100Gbps+ attacks which are not uncommon against major providers.
Most of the time we don't actually operate switches of our own, and our setup is simple enough where we don't do a ton of automation around the config where we do. What we do automate end-to-end is our BGP: our prefixes are actually announced from servers (using exabgp), not routers -- and our config there is managed by ansible, plus some real-time automation around community strings.
A simple example might help. Imagine you have a node in California and a node in NY, and a user in NJ. Let's assume for now that geographic proximity is actually a good arbiter for performance. If we make a bad routing decision and send your user to the CA node, we're adding a lot of overhead in their session with your application: every TCP packet (e.g., every HTTP request) takes a long round trip. Even if we spit out a DNS answer really, really fast, if it's the wrong answer, the user has a bad experience.
It's actually even worse than that, because the user isn't the one directly doing the DNS query: there's an intermediary DNS resolver, which will cache the response. If the "wrong" response gets cached, then every user of that resolver will get that wrong response until the cache expires. So not only have we made the original requester's experience bad, but we've negatively impacted every other user of your application that's leveraging the same resolver.
Time to first byte is a combination of a lot of things, and if you're in more than one datacenter, it doesn't much matter how fast you spit out a DNS response if you're giving the wrong one and impacting the rest of the session going forward.
Doesn't mean you shouldn't expect the best of both worlds: sending the user to the "best" endpoint, fast.
On the filter chain question: typical/canonical approach to do any kind of decision making in a DNS system is to add some new proprietary record type, like a geo record or a health checking record. That's kind of the natural thing to do in DNS at first glance.
But if you want to get any kind of complex routing behavior, you're going to need an awful lot of different record types implementing those different behaviors, and what you end up with behind the scenes is what I often call a spaghetti of different DNS records all pointing at each other in some kind of big decision tree -- maybe a geo record, pointing at a bunch of health checking records, pointing at a bunch of CNAMEs, pointing at a bunch of A records. This quickly becomes unmanageable, and every new kind of routing you want to do results in another layer in the decision tree, all to resolve a single hostname.
The Filter Chain is a way to collapse all that down to something much more manageable and performant by bringing all the context into one place and thinking of routing as a collection of simple actions that you're taking on some input data (answers you could give, and details about those answers).
Certainly a huge drop-off. There is a lot of noise out there -- a lot of "resolvers" doing very few queries, whether because they have very few users, or because they're one-off special purpose resolvers. Vast majority of the internet is really serviced by a few thousand major DNS resolvers. Obviously Google/OpenDNS are chief among those, but Comcast and other major ISPs account for most of the volume, as you'd probably expect.
Hi -- glad you like the post. /32s are individual DNS resolver IPs we observed in each country from among the sample we took (half a percent of the traffic we saw during a 24h period).
The >10Q column is how many of those resolvers we saw that did >10Q from among our sample -- you can see there are many fewer resolvers that are doing even moderate amounts of traffic.
Kris here. Thanks Chris for putting this together, it was a really fun set of questions. I think a lot of what makes NSONE different is the philosophy behind the tech. We're not a DNS company, we're a traffic management company and we think DNS is a great place to do traffic management. Hope that comes across in this article.
At 2M queries/mo our list pricing is $8 (https://nsone.net/support/billing/). Indeed our Biz Plan is $200 -- but that includes 24/7 support, 25M queries, and other bells & whistles. If you just need a little more volume, stick with the startup plan.
That said, indeed, in general we're more expensive than R53, Rage4, and various others, because of the depth of capabilities of the platform and the quality of the support we provide. But we're also pretty unique in giving away 1M queries free with no feature locking, so our most advanced stuff is available to everyone. If you're just after basic geo-routing at the lowest cost, we're happy to help but you may find other services cheaper; but if you intend to go beyond geo to complex failover arrangements, load shedding, weighting/stickiness, network-based fencing, etc, then do some tinkering with NSONE and let us know what you think. Feedback is always good. :)