Efficient Adaptive Read-Ahead in Log Structured Storage

A computer program product comprising a computer readable storage medium having a computer readable program stored therein, wherein the computer readable program, when executed on a computing device, causes the computing device to implement a read-ahead manager for adaptive read-ahead in log structured storage, wherein the computer readable program causes the computing system to:

• determine, by the read-ahead manager, a probability value P representing a probability to read into cache a temporal environment for a front-end read for a given segment in user space in a log structured storage, wherein determining the probability value comprises:
• responsive to reading a plurality of front-end reads of records for a given segment of user space in a log structured storage, determining, by the read-ahead manager, a cache hit ratio associated with the given segment;
• responsive to determining the cache hit ratio is increasing, increasing, by the read-ahead manager, the probability value P;
• and responsive to determining the cache hit ratio is decreasing, decreasing, by the read-ahead manager, the probability value P, wherein determining the cache hit ratio is increasing comprises determining the following:
• H>T×P,
• wherein H is the cache hit ratio, wherein T is a predetermined cache hit threshold;
• and
• responsive to performing a front-end read of a record of the given segment in the log structured storage, perform, by the read-ahead manager, pre-fetch of the temporal environment for the record with probability P.

An apparatus comprising:

• at least one processor;
• and
• at least one memory coupled to the at least one processor, wherein the at least one memory comprises instructions which, when executed by the at least one processor, cause the at least one processor to implement a read-ahead manager for adaptive read-ahead in log structured storage, wherein the instructions cause the at least one processor to:
• determine, by the read-ahead manager, a probability value P representing a probability to read into cache a temporal environment for a front-end read for a given segment in user space in a log structured storage, where determining the probability value comprises:
• responsive to reading a plurality of front-end reads of records for a given segment of user space in a log structured storage, determining, by the read-ahead manager, a cache hit ratio associated with the given segment;
• responsive to determining the cache hit ratio is increasing, increasing, by the read-ahead manager, the probability value P;
• and responsive to determining the cache hit ratio is decreasing, decreasing, by the read-ahead manager, the probability value P, wherein determining the cache hit ratio is increasing comprises determining the following:
• H>T×P,
• wherein H is the cache hit ratio, wherein T is a predetermined cache hit threshold;
• and
• responsive to performing a front-end read of a record of the given segment in the log structured storage, perform, by the read-ahead manager, pre-fetch of the temporal environment for the record with probability P.