Multi-level memory compression

A method comprising:

• selecting, by one or more processors, a page of data that is compressed by a first compression algorithm, wherein the compressed page of data is stored in a first one of a plurality of memory blocks;
• decompressing the page of data by the first compression algorithm, resulting in a decompressed page of data;
• storing the decompressed page of data into an uncompressed memory area;
• combining the first memory block with one or more of the plurality of memory blocks that are unallocated, resulting in a larger memory block;
• identifying one of a plurality of pre-defined memory block sizes corresponding to the larger memory block;
• updating one of a plurality of linked lists corresponding to the identified pre-defined memory block size to indicate that the larger memory block is unallocated;
• recompressing, by one of the processors, the page of data using a second compression algorithm in response to determining that the compressed page of data is utilized less than a utilization threshold, the recompressing resulting in a recompressed page of data;
• determining a size of the recompressed page of data;
• selecting one of the plurality of pre-defined memory block sizes based upon the size of the recompressed page of data, wherein each of the plurality of pre-defined memory block sizes corresponds to a size of a memory block located in memory;
• identifying, in an index array, one of a plurality of indexes that is assigned to the selected pre-defined memory block size;
• determining whether the identified index includes a pointer that points to one of the plurality of linked lists comprising an address location of a second one of the plurality of memory blocks that corresponds to the selected pre-defined memory block size;
• and
• in response to determining that the identified index includes the pointer, storing the recompressed page of data in the second memory block.