Redundancy built on top of redundancy at every layer of the infrastructure stack
If you look at the domain name delegations of several of our largest Dynect DNS customers you will see a common trend. Redundancy. We see many customers running their own hidden master setup, adding their own name servers (NS) into their delegation, or using a multi-provider approach for secondary diversity. Clients work with us to maintain and gain greater traffic control of their DNS and corresponding content (DNS GUI and DNS APIs), while leveraging our global footprint; but even we don’t rely on singular infrastructure providers behind the scenes.
We’ve been following a trend for about a year now, not just about outsourcing services to focused specialists, but maybe to two or three of them. Clients are no longer using one provider for EVERYTHING eitherâ€¦ that’s a faux pas. There is simply too much to lose with too much money being made online (even more during the holiday’s) to have all your eggs in one basket, or in the spirit of trick or treating, all your candy in one pillowcase.
This redundancy movement does not only apply to DNS hosting; it’s application and content delivery overall: origin server setups with diverse collocation (Telx, Equinix, Colospace), managed hosting (Contegix, Pier1, Claranet, Rackspace), cloud providers (GoGrid, Amazon, Rackspace, Joyent), CDN (Akamai, Limelight, Edgecast, MaxCDN), transit (NTT, Level3, Tata)… a multi-provider approach is upon us.
For sure, a big part of our value proposition and why customers choose to outsource their DNS to Dyn, is to gain increased redundancy. On our Dynect Platform for managed DNS and GSLB, each customer delegates to NS 1-4 with NS1/NS3 and NS2/NS4 actually serving laser fast DNS queries from two entirely different DNS constellations or networks. NS1/NS3 make up an 11 location IP Anycast network collocated on our own racks with NS2/NS4 (six PoPs) taking advantage of rich performing CDN, Edgecast Networks. In essence jumping onto the Dynect Platform is almost like jumping on two highly optimized DNS networks. Our overall utilization to capacity ratio is unheard of in the DNS industry.
Customers are also utilizing our advanced features on Dynect to achieve redundancy at some of the other layers of the infrastructure setup we mentioned above. Whether it’s to load balance traffic geographically (keep eyeballs in the US seeing US content and eyeballs in Europe seeing EU content), to simply disperse load across several IPs evenly, to failover to backup or active/active setups in moments of overload disaster, to bleed over CDN traffic for performance and cost control, or to balance between origin servers and globally distributed caching instances… there isn’t a use case or setup we haven’t encountered.
Be sure to check out and sign up: Startup or SMB managed DNS services or Enterprise Anycast DNS and be smart with the rest of your infrastructure rollout. Redundancy has pushed aside the simple “Should we outsource?” debate and is a must.