Techniques for Improving Cache Effectiveness of Caches with Low User Population

Abstract

Techniques for improving cache effectiveness in areas with low user population are provided. In one aspect, a computer-based method for managing user traffic in a cellular network using proxy caches is provided. The method includes the following steps. A group of the proxy caches that has correlated user traffic is identified. Cache requests for each of the proxy caches in the group are observed. One or more patterns are found in the cache requests. A prediction is made as to which items will be requested from the proxy caches in the group in the future based on the one or more patterns found in the cache requests. The proxy caches in the group are pre-populated with the items.

1 . A computer-based method for managing user traffic in a cellular network using proxy caches, the method comprising the steps of:
 identifying a group of the proxy caches that has correlated user traffic, wherein the correlated user traffic takes multiple paths through the cellular network, wherein the cellular network is configured such that each of the paths goes through only one of the proxy caches in the group, and wherein at least one of the paths goes through a different proxy cache in the group from one or more other of the paths;   observing cache requests for each of the proxy caches in the group;   finding one or more patterns in the cache requests;   predicting which items will be requested from the proxy caches in the group in the future based on the one or more patterns found in the cache requests; and   pre-populating the proxy caches in the group with the items.   
     
     
         2 . The method of  claim 1 , wherein the cellular network comprises multiple base stations connected to at least one radio network controller, and wherein one of the proxy caches is present at each of the base stations. 
     
     
         3 . The method of  claim 2 , wherein the step of identifying the group of the proxy caches that has correlated user traffic comprises the steps of:
 identifying the proxy caches that are present at the base stations physically located within a given geographical area, wherein the proxy caches that are present at the base stations physically located within the given geographical area are considered to have correlated user traffic; and   grouping the proxy caches that are present at the base stations physically located within the given geographical area.   
     
     
         4 . The method of  claim 3 , further comprising the step of:
 defining the given geographical area based on pre-existing geographical boundaries.   
     
     
         5 . The method of  claim 4 , wherein the pre-existing geographical boundaries comprise pre-existing geographical boundaries defining towns or cities. 
     
     
         6 . The method of  claim 3 , further comprising the step of:
 defining the given geographical area based on user demographics, wherein the given geographical area comprises users of a certain demographic.   
     
     
         7 . The method of  claim 3 , further comprising the step of:
 defining the given geographical area based on physical proximity of the base stations within the given geographical area to a particular location.   
     
     
         8 . The method of  claim 7 , wherein the given geographical area is defined based on the base stations within the given geographical area being located within a certain distance from a particular type of landmark. 
     
     
         9 . The method of  claim 8 , wherein the landmark comprises a college or a university. 
     
     
         10 . The method of  claim 8 , wherein the landmark comprises an exit on a highway. 
     
     
         11 . The method of  claim 3 , wherein the given geographical area comprises multiple areas two or more of which are not physically continuous with one another. 
     
     
         12 . The method of  claim 1 , wherein the step of observing the cache requests for each of the proxy caches in the group is performed using one or more cache trace collectors. 
     
     
         13 . The method of  claim 12 , wherein one of the cache trace collectors is run at each of the proxy caches. 
     
     
         14 . The method of  claim 2 , wherein the step of finding the one or more patterns in the cache requests and the step of predicting which items will be requested from the proxy caches in the group in the future based on the one or more patterns found in the cache requests are performed using a cache analyzer. 
     
     
         15 . The method of  claim 14 , wherein the cache analyzer is run at the at least one radio network controller. 
     
     
         16 . The method of  claim 1 , further comprising the step of:
 assigning weights to each of the items in each of the proxy caches in the group which indicate how many users requested the items.   
     
     
         17 . The method of  claim 1 , wherein the step of finding one or more patterns in the cache requests comprises the step of:
 identifying a popularity of the items in each of the proxy caches in the group based on a number of times the items were requested from each of the proxy caches in the group.   
     
     
         18 . The method of  claim 17 , further comprising the step of:
 pre-populating the proxy caches in the group with the items that are most popular in each of the proxy caches in the group, wherein the items that are most popular in a given one of the proxy caches are those items requested more than a predetermined number of times from the given proxy cache.   
     
     
         19 . The method of  claim 18 , wherein the step of pre-populating the proxy caches in the group is performed using a cache populator. 
     
     
         20 . The method of  claim 1 , wherein the step of finding one or more patterns in the cache requests comprises the step of:
 determining a probability of one the items being requested from a given one of the proxy caches in the group within a given period of time after another one of the items has been requested from the given proxy cache.   
     
     
         21 . The method of  claim 1 , wherein the step of finding one or more patterns in the cache requests comprises the step of:
 determining a probability that one of the items, once requested from one of the proxy caches in the group, will also be accessed by another one of the caches in the group within a certain period of time.