Cache performance on a storage system

Patent metadata

Field	Value
Publication number	US-10809918-B2
Application number	US-201916263541-A
Country	US
Kind code	B2
Filing date	Jan 31, 2019
Priority date	Jan 31, 2019
Publication date	Oct 20, 2020
Grant date	Oct 20, 2020

Abstract

Options for handling write operations may be selected based on a determined probability that a read operation for a portion of data will occur while the data portion is still in an I/O cache as a result of a write operation. As used herein, a “read-after-write event (“RAW”) is an occurrence of a read operation for a portion of data while the data portion is still in an I/O cache as a result of a write operation. The probability of a RAW may be determined by applying Bayesian inference, and may include applying exponential smoothing to calculations made on historical I/O information so that more recent I/O events have greater weight in determining RAW probability. Based on the determined RAW probability, write data may either be journaled in a write journal or written to a cache slot and de-staged to a physical storage device as part of write-in-place processing.

First claim

What is claimed is:

1. A method of managing an input/output (I/O) cache on a storage system, comprising:
for at least a first storage device of the storage system, determining a probability that a read operation for a portion of data will be received while the portion of data resides in the I/O cache as a result of a write operation for the portion of data;
receiving a write request for the at least first storage device, the write request specifying a write operation for a first portion of data and an address range within the at least first storage device; and
determining based at least in part on the probability whether to: to record a journal entry for the first portion of data in a journal, or write to the I/O cache as part of writing in place.

2. The method of claim 1 , wherein determining the probability includes applying exponential smoothing to historical read and write information so that more recent historical read and write information has greater weight in determining the probability.

3. The method of claim 1 , wherein determining based at least in part on the probability includes:
comparing the probability to a predefined probability threshold;
if the probability is less than or equal to the predefined probability threshold, determining to record the journal entry; and
if the probability is greater than the predefined probability threshold, determining to perform the write to the I/O cache as part of writing in place.

4. The method of claim 1 , further comprising:
if it is determined to record the journal entry, applying a hash function to the address range to produce a hash value,
wherein the journal corresponds to the hash value.

5. The method of claim 4 , further comprising:
storing the hash value in an entry of a hash table, the entry specifying a reference to the journal.

6. The method of claim 1 , wherein the storage system includes one or more journals, the method further comprising:
receiving a read request specifying a read operation for an address range of a storage device; and
determining that either: there is not a journal entry for the address range of the storage device in the one or more journals or there may be a journal entry for the address range of the storage device in a first of the one or more journals.

7. The method of claim 6 , further comprising:
if it determined that there may be a journal entry in the first journal for the address range of the storage device, playing back the first journal to determine if there is a journal entry.

8. A storage system comprising:
an I/O cache;
at least a first storage device;
one or more processors; and
memory comprising code stored thereon that, when executed, performs a method of managing the I/O cache comprising:
for the at least first storage device, determining a probability that a read operation for a portion of data will be received while the portion of data resides in the I/O cache as a result of a write operation for the portion of data;
receiving a write request for the at least first storage device, the write request specifying a write operation for a first portion of data and an address range within the at least first storage device; and
determining based at least in part on the probability whether to: to record a journal entry for the first portion of data in a journal, or write to the I/O cache as part of writing in place.

9. The system of claim 8 , wherein determining the probability includes applying exponential smoothing to historical read and write information so that more recent historical read and write information has greater weight in determining the probability.

10. The system of claim 8 , wherein determining based at least in part on the probability includes:
comparing the probability to a predefined probability threshold;
if the probability is less than or equal to the predefined probability threshold, determining to record the journal entry; and
if the probability is greater than the predefined probability threshold, determining to perform the write to the I/O cache as part of writing in place.

11. The system of claim 8 , the method further comprising:
if it is determined to record the journal entry, applying a hash function to the address range to produce a hash value,
wherein the journal corresponds to the hash value.

12. The system of claim 11 , the method further comprising:
storing the hash value in an entry of a hash table, the entry specifying a reference to the journal.

13. The system of claim 8 , wherein the storage system includes one or more journals, the method further comprising:
receiving a read request specifying a read operation for an address range of a storage device; and
determining that either: there is not a journal entry for the address range of the storage device in the one or more journals; or there may be a journal entry for the address range of the storage device in a first of the one or more journals.

14. The system of claim 13 , the method further comprising:
if it determined that there may be a journal entry in the first journal for the address range of the storage device, playing back the first journal to determine if there is a journal entry.

15. One or more non-transitory computer-readable media, the computer-readable media having software stored thereon for managing an I/O cache of a storage system, the software comprising:
executable code that determines, for at least a first storage device of the storage system, a probability that a read operation for a portion of data will be received while the portion of data resides in the I/O cache as a result of a write operation for the portion of data;
executable code that receives a write request for the at least first storage device, the write request specifying a write operation for a first portion of data and an address range within the at least first storage device; and
executable code that determines based at least in part on the probability whether to: to record a journal entry for the first portion of data in a journal, or write to the I/O cache as part of writing in place.

16. The one or more non-transitory computer-readable media of claim 15 , wherein the executable code that determines the probability includes executable code that applies exponential smoothing to historical read and write information so that more recent historical read and write information has greater weight in determining the probability.

17. The one or more non-transitory computer-readable media of claim 15 , wherein the executable code that determines based at least in part on the probability includes:
executable code that compares the probability to a predefined probability threshold;
executable code that determines, if the probability is less than or equal to the predefined probability threshold, to record the journal entry; and
executable code that determines, if the probability is greater than the predefined probability threshold, to perform the to the I/O cache as part of writing in place.

18. The one or more non-transitory computer-readable media of claim 15 , the software further comprising:
executable code that applies, if it is determined to record the journal entry, a hash function to the address range to produce a hash value,
wherein the journal corresponds to the hash value.

19. The one or more non-transitory computer-readable media of claim 18 , the software further comprising:
executable code that stores the hash value in an entry of a hash table, the entry specifying a reference to the journal.

Assignees

• Emc Ip Holding Co Llc

Inventors

• Martin Owen
• Halstead Mark J
• Smith-Vaniz William R

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N5/046

  Forward inferencing; Production systems · CPC title

• G06F3/0689

  Disk arrays, e.g. RAID, JBOD · CPC title

• G06F3/061 Primary

  Improving I/O performance · CPC title

• G06F3/0656

  Data buffering arrangements · CPC title