Abstract | ||
---|---|---|
With the technology scaling of solid state drives (SSDs), data reliability has to be ensured by advanced Low-Density Parity-Check (LDPC) codes. As LDPC reads often take high latency, read performance involved in traditional garbage collection, called external GC, is largely decreased, which affects system performance. GC can choose not to correct data with LD-PC codes but just copy valid pages into registers of NAND flash, which we call internal GC. Although this method avoids long latency in LDPC reads, the data integrity is sacrificed because of error aggregation in valid pages. To balance performance and reliability, this paper proposes a new garbage collection method, named as LDPC-aware Garbage Collection (LaGC), to selectively choose blocks with high error rates to perform external GC. These blocks are often close to uncorrectable error rates, which take a small portion during a long flash life period. Thus, most of blocks perform internal GC and keep advantage of low read latency. Experimental results show that LaGC can reduce 78% of read latency in external GC and achieve about 2% system performance improvement. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1109/NVMSA.2017.8064481 | 2017 IEEE 6th Non-Volatile Memory Systems and Applications Symposium (NVMSA) |
Keywords | Field | DocType |
LDPC-aware garbage collection,solid state drives,data reliability,read performance,internal GC,data integrity,garbage collection method,LDPC-aware Garbage Collection,SSD performance,LDPC codes,system performance improvement,Low-Density Parity codes,external GC,NAND flash,error aggregation | Low-density parity-check code,Computer science,Latency (engineering),Data reliability,NAND gate,Data integrity,Garbage collection,Decoding methods,Computer hardware,Performance improvement,Embedded system | Conference |
ISSN | ISBN | Citations |
2575-2561 | 978-1-5386-1769-4 | 0 |
PageRank | References | Authors |
0.34 | 6 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yajuan Du | 1 | 8 | 5.21 |
Yunpei Jia | 2 | 0 | 0.34 |
Meng Zhang | 3 | 16 | 5.23 |
Jun Zeng | 4 | 3 | 2.81 |
Chun Jason Xue | 5 | 1616 | 140.95 |