Latency reduction via adaptive speculative preconnection

• US 9,083,583 B1
• Filed: 07/01/2011
• Issued: 07/14/2015
• Est. Priority Date: 07/01/2011
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Roskind, James, Belshe, Michael
• Primary Examiner(s): Vu, Viet
• Assistant Examiner(s): Emdadi, Keyvan
• Application Number:
  US13/175,715
• Time in Patent Office: 1,474 Days
• Field of Search: 709/217, 709/219, 709/227, 709/228, 709/229
• US Class Current: 1/1
• CPC Class Codes: H04L 67/02 based on web technology, e.g. hypertext transfer protocol [HTTP]
• CPC Class Codes: H04L 67/5681 Pre-fetching or pre-delivering data based on network characteristics

Abstract

Methods, systems, and computer program products for reducing web latency using adaptive speculative preconnection are provided. Exemplary methods, systems, and computer program products for reducing web latency include initiating a connection to a first URL hostname, and obtaining speculative information associated with the first URL hostname. The speculative information includes a likelihood of connecting to a second URL hostname as a result of initiating a connection to the first URL hostname. The methods, systems, and computer program products also include determining whether to preconnect to the second URL hostname based on the speculative information. The methods, systems, and computer program products further include providing an instruction to preconnect to the second URL hostname when a determination to preconnect is made.

Claims

1. A computing device-implemented method for reducing web page load time comprising:

initiating a connection to a first uniform resource locator (URL) host;

obtaining speculative information associated with the first URL host, wherein the speculative information comprises a likelihood of connecting to a second URL host as a result of initiating a connection to the first URL host;

determining, using a computing device, whether to preconnect to the second URL host based on the speculative information;

providing an instruction to preconnect to the second URL host when a determination to preconnect is made; and

preconnecting to the second URL host to establish a preconnection when an instruction to preconnect is provided, wherein the preconnecting comprises;

transmitting a synchronization (SYN) packet to the second URL host;

receiving an acknowledgement (ACK) packet in response to the transmitting the SYN packet; and

waiting for an instruction to load content located at the second URL host before transmitting a content request to the second URL host using the preconnection, wherein the preconnection is ready to acquire content from the second URL host when content associated with the first URL host is received over the connection.

2. The method of claim 1, wherein the first URL host and the second URL host have a common domain name.

3. The method of claim 1, wherein the first URL host has a domain name different from the second URL host.

4. The method of claim 1, wherein initiating a connection to a first URL host comprises automatically providing an instruction to preconnect to the second URL host.

5. The method of claim 1, wherein obtaining speculative information comprises receiving the speculative information from a remote site.

6. The method of claim 1, wherein obtaining speculative information comprises obtaining a probability of connecting to the second URL host, an estimate of a number of connections needed to acquire all required subresources of the second URL host, or vectors of state.

7. The method of claim 1, wherein obtaining speculative information comprises consulting a speculative database, the method further comprising:

monitoring requests for resources associated with URL hosts;

estimating the likelihood of connecting to the second URL host based on the monitoring, wherein the likelihood comprises a probability of connecting to the second URL host, an estimate of a number of connections needed to acquire all required subresources of the second URL host, or vectors of state; and

storing a result of the estimating in the speculative database.

8. The method of claim 7, further comprising:

updating an estimate of the likelihood of connecting to the second URL host, wherein the updating comprises reducing the estimate by dividing the estimate by a number; and

storing a result of the updating in the speculative database.

9. The method of claim 7, wherein the estimating is based at least in part on at least one of historical estimates, remote site estimates, and the monitoring.

10. The method of claim 7, wherein the estimating depends on the hostname of the first URL host and the estimating is aggregated across all observed paths of the first URL host.

11. The method of claim 7, wherein the estimating depends on a hostname and part of a path of the first URL hostname and the estimating is aggregated across the first URL hostname and part of the path of the first URL hostname.

12. The method of claim 1, wherein the preconnecting further comprises transmitting an ACK in response to the receiving an acknowledgement.

13. The method of claim 1, wherein initiating a connection to a first URL host comprises performing a Domain Name System (DNS) lookup, connecting to the first URL host, and acquiring content associated with the first URL host such that the initiating a connection to a first URL host occurs during a time period that overlaps with preconnecting to the second URL host.

14. The method of claim 1, wherein initiating a connection to a first URL host comprises initiating a secure connection, the method further comprising:

receiving a session key for the secure connection to the first URL host; and

using the session key to communicate with the second URL host.

15. A system for reducing web page load time comprising:

a connector, implemented with a computing device, to initiate a connection to a first uniform resource locator (URL) host;

an adapter configured to obtain speculative information associated with the first URL host, wherein the speculative information comprises a likelihood of connecting to a second URL host as a result of initiating a connection to the first URL host; and

a speculator configured to

determine whether to preconnect to the second URL host based on the speculative information, and

provide an instruction to preconnect to the second URL hostname when a determination to preconnect is made;

a preconnector configured to preconnect to the second URL host to establish a preconnection when an instruction to preconnect to the second URL host is provided, wherein the preconnector transmits a synchronization (SYN) packet, receives an acknowledgement (ACK) packet in response to transmitting the SYN packet, and waits for an instruction to load content located at the second URL host before transmitting a content request to the second URL host using the preconnection, wherein the preconnection is ready to acquire content from the second URL host when content associated with the first URL host is received over the connection.

16. The system of claim 15, further comprising:

a request monitor, implemented with a computing device, to monitor requests for resources associated with URL hosts; and

the speculator further configured to

estimate the likelihood of connecting to the second URL host based on the monitoring, wherein the likelihood comprises a probability of connecting to the second URL host, an estimate of a number of connections needed to acquire all required subresources of the second URL host, or vectors of state, and

store a result of the estimate in the speculative database.

17. The system of claim 16, wherein when the speculator estimates a likelihood of connecting to the second URL host, the speculator is based at least in part on at least one of historical estimates, remote site estimates, and the monitoring.

18. The system of claim 15, wherein when the adapter obtains the speculative information, the adapter obtains a probability of connecting to the second URL host, an estimate of a number of connections needed to acquire all required subresources of the second URL host, or vectors of state.

19. The system of claim 15, wherein when the connector initiates a connection to the first URL host, the connector performs a Domain Name System (DNS) lookup, connects to the first URL host, and acquires content associated with the first URL host such that initiating the connection to the first URL host occurs during a time period that overlaps when the preconnector preconnects to the second URL host.

20. The system of claim 15, wherein the connector initiates a secure connection to the first URL host and receives a session key for the secure connection to the first URL host, and the session key is used to communicate with the second URL host.

21. A non-transitory computer-program product comprising a computer-usable medium having computer program logic recorded thereon for enabling a processor to reduce web page load time, the computer program logic comprising computer readable program code that enables a processor to perform operations to:

initiate a connection to a first uniform resource locator (URL) host;

to obtain speculative information associated with the first URL host, wherein the speculative information comprises a likelihood of connecting to a second URL host as a result of initiating the connection to the first URL host;

determine whether to preconnect to the second URL host based on the speculative information;

provide an instruction to preconnect to the second URL host when a determination to preconnect to the second URL host is made; and

preconnect to the second URL host to establish a preconnection when an instruction to preconnect is provided, wherein the preconnect comprises operations to;

transmit a synchronization (SYN) packet to the second URL host;

receive an acknowledgement (ACK) packet in response to transmission of the SYN packet; and

wait for an instruction to load content located at the second URL host before transmitting a content request to the second URL host using the preconnection, wherein the preconnection is ready to acquire content from the second URL host when content associated with the first URL host is received over the connection.

22. The non-transitory computer program product of claim 21, further comprising operations to:

monitor requests for resources associated with URL hosts;

estimate the likelihood of connecting to the second URL host based on the monitoring, wherein the likelihood comprises a probability of connecting to the second URL host, an estimate of a number of connections needed to acquire all required subresources of the second URL host, or vectors of state; and

store a result of the estimating in the speculative database.

23. The non-transitory computer program product of claim 21, wherein the preconnect includes operations to perform a DNS lookup, connect to the first URL host, and acquire content associated with the first URL host,

wherein the initiating the connection to the first URL host occurs during a time period that overlaps with preconnecting to the second URL host.