Home > Storage > PowerScale (Isilon) > Product Documentation > Storage (general) > PowerScale OneFS SmartFlash > Cache level comparison
L2 cache is typically more valuable than L1 cache because a hit avoids a higher latency operation. An L1 cache hit avoids a back-end round trip to fetch the data, whereas an L2 cache hit avoids a SATA disk seek in the worst case. This difference is dramatic in both relative and absolute terms. For SATA drives, an L2 cache miss is two orders of magnitude above a hit compared to one for L1 cache. A single back-end round trip is typically a small portion of a full front-end operation.
Cache | Hit | Miss |
L1 | 10 µs | L2 |
L2 | 100 µs | L3 (or Hard Disk) |
L3 | 200 µs | Hard Disk |
Hard Disk | 1 ms–10 ms | x |
L2 cache is also preferable because it is accessible to all nodes. Assuming a workflow with any overlap among nodes, it is preferable to have the cluster’s DRAM holding L2 cache data rather than L1 cache data. In L2 cache, a given data block is only cached once and invalidated much less frequently. This efficiency is why storage nodes are configured with a drop-behind policy on file data. Nodes without disks will not drop behind because there is no L2 cache data to be cached.
Metadata in L1 cache is not dropped behind because it is accessed so frequently. For example, when streaming a file, the inode is accessed for every new protection group (PG) read, while each data block is only accessed once. Metadata is also multiple steps in the latency path.