FAULT TOLERANCE APPROACHES FOR DNS SERVER FAILURES

• US 20100332680A1
• Filed: 10/29/2009
• Published: 12/30/2010
• Est. Priority Date: 06/24/2009
• Status: Active Grant

• Current Assignee: Avago Technologies International Sales Pte Limited (Broadcom, Inc.)
• Original Assignee: Broadcom Corporation (Broadcom, Inc.)
• Inventors: Nechamkin, Eugene, Djurovic, Dragan, Anderson, Sam, Shoghli, Jamshid
• Granted Patent: US 8,935,428 B2
• Time in Patent Office: Days
• US Class Current: 709/245
• CPC Class Codes: H04L 61/4511 using domain name system [DNS]
• CPC Class Codes: H04L 61/58 Caching of addresses or names

Abstract

Techniques are provided for handling failures of DNS (domain name system) servers to respond to DNS queries. A DNS resolver is configured to resolve domain names, and includes a time-to-live (TTL)-based cache, a negative cache, and a long term store cache. The TTL-based cache is configured to temporarily store domain names with resolved IP addresses. The negative cache is configured to store negative entries that include information indicating domain names that were failed to be resolved. The long term store cache is configured to store domain names with resolved IP address for an indefinite time period. The caches are accessed in a manner that enables fewer DNS query retries to be performed when a DNS server is non-responsive, to reduce delays and network traffic. Furthermore, the DNS resolver may reduce a number of DNS queries performed the longer the DNS server stays non-responsive.

Claims

1. A method in a first communication device, comprising:

receiving a request to resolve an address for a domain name corresponding to a second communication device;

accessing a local time-to-live (TTL)-based cache for an address corresponding to the domain name;

accessing a local negative cache for a negative entry corresponding to the domain name if the address is not present in the TTL-based cache;

accessing a local long term store cache for the address corresponding to the domain name if the negative entry is present in the TTL-based cache; and

enabling a communication signal to be transmitted to the second communication device according to the address if the address is present in at least one of the TTL-based cache or the long term store cache.

2. The method of claim 1, wherein if the address is not present in the TTL-based cache and the negative entry is not present in the negative cache, performing:

transmitting a DNS (domain name system) query to a DNS server to request the address corresponding to the domain name;

storing the address in the TTL-based cache and in the long term store cache if the address is received from the DNS server in response to the DNS query;

enabling the communication signal to be transmitted to the second communication device according to the address if the address is received from the DNS server in response to the DNS query; and

storing a negative entry in the negative cache, accessing the long term store cache for the address corresponding to the domain name, and enabling the communication signal to be transmitted to the second communication device using the address accessed in the long term store cache if the DNS server provides an error message in response to the DNS query.

3. The method of claim 2, wherein if the DNS server does not respond to the transmitted DNS query, performing

executing a back-off retry operation having a retry count value that defines a number of retry attempts for transmitting the DNS query to the DNS server;

determining that a DNS failure has occurred if a number of retry attempts corresponding to the retry count value is performed without receiving the requested address from the DNS server; and

entering a failure mode if a DNS failure is determined to have occurred.

4. The method of claim 3, wherein said executing a back-off retry operation comprises:

waiting a length of time based on a predetermined time out value;

transmitting the DNS query to the DNS server after the length of time expires; and

waiting an increased length of time and retransmitting the DNS query to the DNS server after the increased length of time each time that the DNS server does not respond to the transmitted DNS query until the retry count value for the back-off retry operation is reached.

5. The method of claim 3, wherein if the address is received from the DNS server in response to a DNS query during the back-off retry operation, performing

storing the address in the TTL-based cache and in the long term store cache;

and

enabling the communication signal to be transmitted to the second communication device.

6. The method of claim 3, wherein if a DNS failure is determined to have occurred, performing;

decrementing the retry count value;

storing a negative entry in the negative cache;

accessing the long term store cache for the address corresponding to the domain name; and

enabling the communication signal to be transmitted to the second communication device according to the address accessed in the long term store cache.

7. The method of claim 6, further comprising:

receiving a second request to resolve an address for the domain name;

accessing the local negative cache for a negative entry corresponding to the domain name;

transmitting a second DNS query to the DNS server to request the address corresponding to the domain name if the negative entry is not present in the negative cache;

re-executing the back-off retry operation having a number of retry attempts indicated by the retry count value if the DNS server does not respond to the transmitted second DNS query; and

decreasing the retry count value, storing a negative entry in the negative cache, accessing the long term store cache for the address corresponding to the domain name, and enabling a second communication signal to be transmitted to the second communication device according to the address accessed in the long term store cache if the DNS server does not respond to the transmitted second DNS query during the re-executed back-off retry operation.

8. The method of claim 7, wherein if the address is received from the DNS server in response to a DNS query during the re-executed back-off retry operation, performing

resetting the retry count value to a predetermined original retry count value;

storing the address in the TTL-based cache and in the long term store cache; and

enabling a second communication signal to be transmitted to the second communication device.

9. A first communication device, comprising:

a DNS (domain name system) resolver configured to process a request to resolve an address for a domain name corresponding to a second communication device, the DNS resolver including

a time-to-live (TTL)-based cache;

a negative cache;

a long term store cache; and

a cache access logic module configured to access the local TTL-based cache for an address corresponding to the domain name;

wherein the cache access logic module is configured to access a local negative cache for a negative entry corresponding to the domain name if the address is not present in the TTL-based cache;

wherein the cache access logic module is configured to access a local long term store cache for the address corresponding to the domain name if the negative entry is present in the TTL-based cache; and

wherein the cache access logic module is configured to enable the communication signal to be transmitted to the second communication device according to the address if the address is present in at least one of the TTL-based cache or the long term store cache.

10. The first communication device of claim 9, wherein if the address is not present in the TTL-based cache and the negative entry is not present in the negative cache, the cache access logic module is configured to transmit a DNS (domain name system) query to a DNS server to request the address corresponding to the domain name, to store the address in the TTL-based cache and in the long term store cache if the address is received from the DNS server in response to the DNS query, and to enable the communication signal to be transmitted to the second communication device according to the address if the address is received from the DNS server in response to the DNS query;

and

wherein the cache access logic module is configured to store a negative entry in the negative cache, to access the long term store cache for the address corresponding to the domain name, and to enable the communication signal to be transmitted to the second communication device using the address accessed in the long term store cache if the DNS server provides an error message in response to the DNS query.

11. The first communication device of claim 10, wherein the cache access logic module includes a back-off retry module;

and

wherein if the DNS server does not respond to the transmitted DNS query, the back-off retry module is configured to execute a back-off retry operation having a predetermined retry count value that defines a number of retry attempts for transmitting the DNS query to the DNS server, and the cache access logic module is configured to enter a failure mode if a number of retry attempts corresponding to the retry count value is performed without receiving the requested address from the DNS server.

12. The first communication device of claim 11, wherein in order to execute the back-off retry operation, the back-off retry module is configured to wait a length of time based on a predetermined time out value, to enable the DNS query to be transmitted to the DNS server after the length of time expires, and to wait an increased length of time and enable the DNS query to be retransmitted to the DNS server after the increased length of time each time that the DNS server does not respond to the transmitted DNS query until the retry count value for the back-off retry operation is reached.

13. The first communication device of claim 11, wherein if the address is received from the DNS server in response to a DNS query during the back-off retry operation, the cache access logic module is configured to store the address in the TTL-based cache and in the long term store cache and to enable the communication signal to be transmitted to the second communication device.

14. The first communication device of claim 11, wherein if the cache access logic module is in failure mode, the back-off retry is configured to decrement the retry count value, and the cache access logic module is configured to store a negative entry in the negative cache, to access the long term store cache for the address corresponding to the domain name, and to enable the communication signal to be transmitted to the second communication device according to the address accessed in the long term store cache.

15. The first communication device of claim 14, wherein if a second request to resolve an address for the domain name is received:

the cache access logic module is configured to access the local negative cache for a negative entry corresponding to the domain name, and to enable a second DNS query to be transmitted to the DNS server to request the address corresponding to the domain name if the negative entry is not present in the negative cache;

the back-off retry module is configured to re-execute the back-off retry operation having a number of retry attempts indicated by the retry count value if the DNS server does not respond to the transmitted second DNS query, and to decrease the retry count value; and

the cache access logic module is configured to store a negative entry in the negative cache, to access the long term store cache for the address corresponding to the domain name, and to enable a second communication signal to be transmitted to the second communication device according to the address accessed in the long term store cache if the DNS server does not respond to the transmitted second DNS query during the re-executed back-off retry operation.

16. The method of claim 15, wherein if the address is received from the DNS server in response to a DNS query during the re-executed back-off retry operation, the back-off retry module is configured to reset the retry count value to a predetermined original retry count value, and the cache access logic module is configured to store the address in the TTL-based cache and in the long term store cache, and to enable a second communication signal to be transmitted to the second communication device.

17. The first communication device of claim 10, wherein the first communication device is coupled to the DNS server through a DOCSIS (data over cable service interface specification) network.

18. The first communication device of claim 9, wherein the first communication device is an internet protocol (IP) telephone.

19. The first communication device of claim 9, wherein the first communication device is a computer and the second communication devices is a web server.

20. The first communication device of claim 9, wherein the communication signal is an instant message.