摘要 |
A method, system, and computer program product are disclosed for managing data in a cache. A first cache memory is provided that includes data. A second cache memory is provided that also includes data in which at least some of the data in the first cache memory is the same as at least some of the data in the second cache memory. In response to a request for data that is stored in both the first and second cache memories, one of the cache memories is chosen in accordance with an access balancing technique. The access balancing technique may include at least one of selection using round robin, and selection based on statistical analysis such as access frequency of the first and second cache memories. First and second access balancing techniques may be used for data accessed from the cache memories, which may comprise disk data and control data, such as data indicating whether data in the caches has been modified and is write pending. |