Method and system for providing on-demand content delivery for an origin server

Abstract

An infrastructure “insurance” mechanism enables a Web site to fail over to a content delivery network (CDN) upon a given occurrence at the site. Upon such occurrence, at least some portion of the site's content is served preferentially from the CDN so that end users that desire the content can still get it, even if the content is not then available from the origin site. In operation, content requests are serviced from the site in the usual manner, e.g., by resolving DNS queries to the site's IP address, until detection of the given occurrence. Thereafter, DNS queries are managed by a CDN dynamic DNS-based request routing mechanism so that such queries are resolved to optimal CDN edge servers. After the event that caused the occurrence has passed, control of the site's DNS may be returned from the CDN back to the origin server's DNS mechanism.

1 . A method to protect a server, wherein data deliverable from the server is associated with a first domain, the first domain identified in a domain name service (DNS) record, comprising:
 in response to receipt of an indication of an occurrence of a given condition associated with the server, using a hardware element to perform an automated rewrite of the DNS record such that that the first domain points to a second domain, the second domain being distinct from the first domain and being associated with a service provider, wherein the given condition is one of: a failure at the server, an occurrence of excess demand at a Web site hosted on the server, a receipt of a request for content that cannot then be served from the server, an occurrence of excess traffic to the Web site originating from a given geography or network, an occurrence of excess latency at the Web site as measured by network agents, and a denial of service attack; and   following termination of the given condition, using the hardware element to update the DNS record such that the first domain no longer points to the second domain.   
     
     
         2 . The method as described in  claim 1  wherein the rewriting of the DNS record associates a canonical name with the first domain. 
     
     
         3 . The method as described in  claim 1  wherein the server is an origin server. 
     
     
         4 . The method as described in  claim 3  wherein the service provider provides delivery of the data on behalf of the origin server. 
     
     
         5 . The method as described in  claim 1  wherein the rewriting occurs in a name service associated with the server. 
     
     
         6 . A system to protect a server, wherein data deliverable from the server is associated with a first domain, the first domain identified in a domain name service (DNS) record, comprising:
 a name service;   a monitor; and   a control routine executing in hardware in response to receipt of an indication from the monitor of an occurrence of a given condition associated with the server, to cause the name service to perform an automated rewrite of the DNS record such that that the first domain points to a second domain, the second domain being distinct from the first domain and being associated with a service provider, wherein the given condition is one of: a failure at the server, an occurrence of excess demand at a Web site hosted on the server, a receipt of a request for content that cannot then be served from the server, an occurrence of excess traffic to the Web site originating from a given geography or network, an occurrence of excess latency at the Web site as measured by network agents, and a denial of service attack; and   the control routine further executing in hardware following termination of the given condition as indicated by the monitor to update the DNS record such that the first domain no longer points to the second domain.   
     
     
         7 . The system as described in  claim 6  wherein the monitor is associated with the service provider.