Is BigHash.Bucket.Slot.Size and BlockCache.Region.Entry.EntryDesc.KeyHash do not need serialize to disk.

[SwordHoly]

As upper pictures shows, BigHash.Bucket.Slot.Size can be calculated by Slot.BucketEntry.KeySize + Slot.BucketEntry.ValueSize, BlockCache.Region.Entry.EntryDesc.KeyHash can be calculated from BlockCache.Region.Entry.EntryValue.Key. So I think it's do not need to serialize BigHash.Bucket.Slot.Size and BlockCache.Region.Entry.EntryDesc.KeyHash, and can save disk space. Especially BigHash is for smallItem, maybe KeySize and ValueSize is jiust 1 byte, the 4 bytes size overhead is not small.

[sunnyszy]

BTW @SwordHoly the slides look very nice. Can you share the full version? I could not find them by search.

[leozzx]

I think this is to reduce the cpu cost during lookup, the disk space is not a top constraint comparing to cpu and memory.
Adding up to you, there's also an 8 bytes keyHash inside the BucketEntry.

[SwordHoly]

I think this is to reduce the cpu cost during lookup, the disk space is not a top constraint comparing to cpu and memory. Adding up to you, there's also an 8 bytes keyHash inside the BucketEntry.

Thanks for replay, i think cpu cost for calucate SlotSize from KeySize and ValueSize is also very small, maybe cachelib do this is just for code hierarchy