Integration of service scaling and service discovery systems

• US 10,182,033 B1
• Filed: 09/19/2016
• Issued: 01/15/2019
• Est. Priority Date: 09/19/2016
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Farhangi, Alireza, Barclay, Christopher, Suresh, Sairam, Rognant, Pierre, Meckley, Sean Henry Lewis
• Primary Examiner(s): Bengzon, Greg C
• Application Number:
  US15/269,822
• Time in Patent Office: 848 Days
• CPC Class Codes: H04L 41/0886 Fully automatic configuration
• CPC Class Codes: H04L 41/0895 Configuration of virtualised networks or elements, e.g. virtualised network function or OpenFlow elements
• CPC Class Codes: H04L 41/40 using virtualisation of network functions or resources, e.g. SDN or NFV entities
• CPC Class Codes: H04L 43/0876 Network utilisation, e.g. volume of load or congestion level
• CPC Class Codes: H04L 43/16 Threshold monitoring
• CPC Class Codes: H04L 43/20 the monitoring system or the monitored elements being virtualised, abstracted or software-defined entities, e.g. SDN or NFV
• CPC Class Codes: H04L 61/4511 using domain name system [DNS]
• CPC Class Codes: H04L 67/1036 Load balancing of requests to servers for services different from user content provisioning, e.g. load balancing across domain name servers

Abstract

Systems and methods are described to enable integrating operation of a service record system with operation of an automatically scaled service hosting system. The service hosting system can maintain a set of servers to provide a network-accessible service, and the service record system can maintain records identifying the set of servers as endpoints for the service. The service hosting system can further modify the number of servers within the set based, for example, on demand. When the service hosting system intends to remove a server from the set, it may notify the service record system. The service record system, in turn, can determine whether any valid records are predicted to exist that identify the to-be-removed server as an endpoint for the service. If such records are predicted to exist, removal of the server can be delayed until those records expire, to prevent errors resulting from client reliance on those records.

Claims

1. A system for managing automatic scaling of a pool of servers based on domain name system (DNS) records associated with the pool of servers, the system comprising:

a hosting system configured with computer executable instructions to manage the pool of servers, wherein the pool of servers includes a plurality of servers collectively configured to implement a network-accessible service, and wherein the hosting system is configured to modify a number of servers within the pool of servers based at least in part on a demand for the network-accessible service; and

a resolver system comprising a processor configured with computer executable instructions that when executed cause the system to;

receive client requests to resolve an identifier of the network-accessible service into a set of network addresses; and

respond to the client requests by providing the DNS records, wherein the DNS records identify network addresses for at least some of the plurality of servers within the pool;

wherein the computer executable instructions, when executed, further cause the resolver system to;

receive a notification that the hosting system intends to remove a first server from the pool of servers;

request that the hosting system delay removal of the first server;

determine a point in time at which no valid DNS records are determined to exist that identify the first server as an endpoint for the network-accessible service, wherein the point in time is determined based at least partly on a time-to-live (TTL) value of the DNS records;

determine that the point in time has occurred and that no valid DNS records exist that identify the first server as an endpoint for the network-accessible service; and

after determining that no valid DNS records to exist that identify the first server as an endpoint for the network-accessible service, transmit instructions to the hosting system to proceed with removal of the first server from the pool of servers.

2. The system of claim 1, wherein the DNS resolver is further configured with computer executable instructions to:

obtain a request to associate the pool of servers with the DNS records; and

transmit a request to the host system to receive notifications regarding modifications to a number of servers within the pool of servers.

3. The system of claim 1, wherein the DNS resolver is configured to determine the point in time at least partly by:

determining a reference time; and

incrementing the reference time with a TTL of the DNS records to result in the point in time.

4. The system of claim 3, wherein the reference time is at least one of a past time at which a DNS record was transmitted to identify the network address of the first server as an endpoint for the network-accessible service or a current time.

5. The system of claim 1, wherein the DNS resolver is further configured with computer executable instructions to:

receive a notification that the hosting system intends to add a second server to the pool of servers; and

modify the DNS records to include a network address of the second server.

6. A computer-implemented method comprising:

obtaining information associating a pool, comprising a plurality of servers configured to implement a network-accessible service, with DNS records identifying endpoints of the network-accessible service;

receiving a notification from a hosting system that a first server, of the plurality of servers, is intended for removal from the pool;

requesting that the hosting system delay removal of the first server;

determining a point in time at which no valid DNS records are forecasted to exist that identify the first server as an endpoint for the network-accessible service;

determining that the point in time has occurred and that no valid DNS records are forecasted to exist that identify the first server as the endpoint for the network-accessible service; and

after determining that no valid DNS records are forecasted to exist that identify the first server as the endpoint for the network-accessible service, transmitting instructions to the hosting system to proceed with removal of the first server from the pool of servers.

7. The computer-implemented method of claim 6 further comprising:

receiving a notification that the hosting system intends to add a second server to the pool of servers; and

modifying the DNS records to include a network address of the second server.

8. The computer-implemented method of claim 6, wherein determining the point in time comprises:

determining a reference time; and

incrementing the reference time with a TTL of the DNS records to result in the point in time.

9. The computer-implemented method of claim 8, wherein the reference time is at least one of a past time at which a DNS record to identify the network address of the first server as an endpoint for the network-accessible service was transmitted or a current time.

10. The computer-implemented method of claim 8, wherein determining the point in time further comprises determining a propagation time for a DNS record to at least one DNS server and incrementing the point in time according to the propagation time.

11. The computer-implemented method of claim 6, wherein validity of DNS records is determined based at least in part on the TTL value.

12. The computer-implemented method of claim 6, wherein the network address is an internet protocol (IP) address.

13. A system comprising:

a data store including information associating a pool, comprising a plurality of servers configured to implement a network-accessible service, with service records identifying endpoints of the network-accessible service; and

a processor configured with computer-executable instructions that when executed cause the system to;

receive a notification that a first server, of the plurality of servers, will become unavailable to provide the network-accessible service;

determine a point in time at which valid service records are not forecasted to exist that identify the first server as an endpoint for the network-accessible service, wherein the point in time is determined based at least partly on a time-to-live (TTL) values of the service records;

determine that the point in time has occurred and that no valid DNS records are forecasted to exist that identify the first server as the endpoint for the network-accessible service; and

after determining that no valid DNS records are forecasted to exist that identify the first server as the endpoint for the network-accessible service, transmit instructions to the hosting system to proceed with rendering the first server unavailable to provide the network-accessible service.

14. The system of claim 13, wherein the computing device is further configured with computer-executable instructions to request that the hosting system delay rendering the first server unavailable to provide the network-accessible service.

15. The system of claim 13, wherein the service records include DNS records.

16. The system of claim 13, wherein the computing device is configured to determine the point in time at least partly by:

determining a reference time; and

incrementing the reference time with a TTL of the DNS records to result in the point in time.

17. The system of claim 16, wherein the reference time is at least one of a past time at which a DNS record was transmitted to identify the network address of the first server as an endpoint for the network-accessible service or a current time.

18. The system of claim 13, wherein the computing device is configured to determine the point in time at least partly by incrementing a reference time with a propagation delay associated with the service records.

19. The system of claim 13, wherein the computing device is configured to:

receive a notification that the hosting system intends to add a second server to the pool of servers; and

modify the service records to identify the second server as an endpoint for the network-accessible service.

20. The system of claim 13, wherein the service records identify the first server as an endpoint for the network-accessible service based at least in part on a network address of the first service.

21. The system of claim 13, wherein the notification that the first server will become unavailable to provide the network-accessible service includes at least one of a notification that the first server will be decommissioned from the pool or a notification that the first server is undergoing maintenance.