Various embodiments relating to a social cache replacement policy are described. The techniques of the present invention disclosed utilize social network properties to guide a cache replacement policy executed by a social networking platform system. In one embodiment, a method is provided for determining a queue location to cache a data item based on a popularity score computed from social network properties. In one embodiment, a method is provided for computing the popularity score by incorporating a user's social network properties and the user's friends' social network properties. In embodiments, the popularity score may be computed using a plurality of social network properties, which may include social network properties associated with (i) the user, (ii) the consumer(s), and/or (iii) the data item(s). In embodiments, a plurality of popularity scores are maintained in a user-score database, where the plurality of popularity scores are periodically updated using historical data.
1. A method performed by a computing system, comprising:
receiving, in a networking platform system, a data item generated by a posting user of the networking platform system;
determining, by the networking platform system, a popularity score associated with the posting user as a function of multiple social properties, the social properties including a login probability that at least one of multiple users logs in to the networking platform system within a specified time subsequent to the data item being generated;
identifying, by the networking platform system, a queue location for the data item based on the popularity score, the queue location being a location in a cache queue of the networking platform system; and
inserting the data item into the cache queue at the identified queue location.
2. The method of claim 1 further comprising:
identifying, within the cache queue, a specified data item having a lowest popularity score; and
removing the specified data item from the cache queue.
3. The method of claim 1 , wherein determining the popularity score further includes:
performing a look-up search in a user-score database by matching the popularity score with the posting user based on a user ID.
4. The method of claim 1 further comprising:
storing multiple popularity scores corresponding to multiple posting users of a networking platform system in a user-score database; and
periodically updating the popularity scores in the user-score database.
5. The method of claim 1 , wherein the social properties are associated with any of the posting user, a consumer of the data item, or the data item, wherein the consumer is one of the users of the networking platform system.
6. The method of claim 1 , wherein the social properties include an interest probability, and wherein determining the popularity score further includes:
determining the interest probability as a function of a number of interest indicators received from a group of the users for past data items posted by the posting user, the group of the users having a social network connection to the posting user in a social network of the posting user, and
computing the popularity score based on the interest probability.
7. The method of claim 6 , wherein determining the interest probability further includes:
determining the interest probability as a function of a number of interest indicators received from a subset of the group of the users for past data items posted by the posting user, the subset of group of the users having a social network connection to both the posting user and the remaining of the group of the users.
8. The method of claim 1 , wherein determining the popularity score further includes:
determining the popularity score as a function of a recentness probability that the data item is likely recent, the recentness probability being one of the social properties includes.
9. The method of claim 8 , wherein determining the recentness probability includes:
determining the recentness probability as a function of a total number of the users having a social network connection to a first user of the users having a social network connection with the posting user.
10. The method of claim 8 , wherein the recentness probability indicates the recentness of the data item relative to multiple data items posted by users having a social network connection to a first user of the users having a social network connection with the posting user.
11. The method of claim 1 , wherein determining the login probability includes:
determining a likelihood that at least one of the users logs in to the networking platform system within a specified time subsequent to the generation of the data item.
12. The method of claim 1 , wherein determining the login probability includes:
determining the login probability of at least one of the users using a login history of the at least one of the users, the login history comprising login information of the at least one of the users for a specified period.
13. The method of claim 1 , wherein determining the queue location includes using a placement algorithm with popularity score, the placement algorithm being at least one of a deterministic scheme or a stochastic scheme.
14. A non-transitory computer-readable storage medium storing computer-readable instructions, comprising:
instructions for selecting a posting user of a networking platform system, the posting user being a user who posts a data item in the networking platform system;
instructions for determining multiple users having an online social connection to the posting user in a social network of the posting user, the group of users authorized to access the data item;
instructions for identifying, by the networking platform system, multiple social properties, the social properties including a login probability that at least one of the users logs in to the networking platform system within a specified time subsequent to the data item being posted, the social properties including a recentness probability that the data item is likely recent; and
instructions for computing, based on the social properties, an access-request probability that is indicative of a likelihood that the data item will be accessed by the at least one of the users within the specified time.
15. The non-transitory computer-readable storage medium of claim 14 further comprising:
instructions for determining an order in which multiple data items are to be stored in a cache queue based on the access-request probabilities associated with the data items.
16. The non-transitory computer-readable storage medium of claim 14 feature comprising:
instructions for storing the access-request probability as a popularity score associated with the posting user in a user-score database.
17. The non-transitory computer-readable storage medium of claim 16 further comprising:
instructions for caching multiple data items into a cache queue by incorporating the popularity score into one or more cache replacement algorithms.
18. The non-transitory computer-readable storage medium of claim 14 , wherein the instructions for computing the access-request probability include:
instructions for computing the access-request probability using historical data stored in a user score database, the historical data containing the social properties; and
instructions for storing, in the user-score database, the access-request probability as a proxy for a popularity score associated with the posting user.
19. The non-transitory computer-readable storage medium of claim 14 further comprising:
instructions for determining a queue location based on the access-request probability associated with the posting user; and
instructions for inserting the data item into a cache queue based on the queue location.
20. A system, comprising:
a processor;
a first module configured to receive a data item generated by a posting user of a networking platform system;
a second module configured to determine a popularity score associated with the posting user as a function of multiple social properties, the social properties including a login probability that at least one of multiple users logs in to the networking platform system within a specified time subsequent to the data item being generated;
a third module configured to identify a queue location for the data item based on the popularity score, the queue location being a location in a cache queue of the networking platform system; and
a fourth module configured to insert the data item into the cache queue at the determined queue location.