Network informed policy creation using dynamically sized windows

Patent metadata

Field	Value
Publication number	US-12007844-B2
Application number	US-202318308038-A
Country	US
Kind code	B2
Filing date	Apr 27, 2023
Priority date	Sep 30, 2021
Publication date	Jun 11, 2024
Grant date	Jun 11, 2024

Abstract

Optimizing backups for a data protection system by determining a size of a dataset to be saved in each backup operation and an available bandwidth in each transfer window of transfer windows for data transfers between a source and destination. An order of the backup operations is defined based on first ordering the backups based on decreasing dataset size and second ordering the transfer windows in order of decreasing bandwidth, and matching the backups to the transfer windows in accordance with these orderings. A dynamic window sizing process determines an initial change between the minimum and maximum bandwidth utilization over a period of time and then iteratively split and consolidate the time blocks until optimal utilization over the time period is reached.

First claim

What is claimed is:

1. A computer-implemented process of determining an optimal window size for backing up data in a series of time windows of a backup operation, comprising:
dividing a backup window of the backup operation into a plurality of slices;
assigning a respective percentage of a minimum and maximum available bandwidth to each slice of the plurality of slices;
iteratively determining which slices that have more than a certain percent difference between the assigned minimum and maximum available bandwidth;
splitting the determined slices into sub-slices into 1-minute, and assigning the minimum and maximum available bandwidth to the sub-slices;
determining which immediately consecutive slices have, when combined, a bandwidth difference that is less than or equal to the certain percent difference to identify consecutive under-utilized slices; and
consolidating the consecutive under-utilized slices together.

2. The method of claim 1 further comprising:
processing all remaining sub-slices for consolidation into consolidated slices; and
assigning the minimum and maximum available bandwidth values to each consolidated slice.

3. The method of claim 2 further comprising computing an overall bandwidth loss, wherein the overall bandwidth loss is an average of minimum and maximum bandwidth losses multiplied by the percentage of time each slice occupies in the entire backup window.

4. The method of claim 3 further comprising:
determining if the total loss is less than a value that represents an optimal value defined by one or more threshold conditions comprising: a total loss is less than a first defined threshold, or an improvement in bandwidth loss less than a second defined threshold value from a prior iteration, or if X is less than a third minimum threshold value; and
proceeding to a cleanup process if the total loss is less than the value, or iterating to a next time slice if the total loss is not less than the value.

5. The method of claim 4 wherein the cleanup process comprises:
consolidating small slices defined as slices less than a defined duration long into larger slices, wherein each small slice is consolidated with one of two consecutive slices that minimizes the bandwidth loss across the two consecutive slices;
updating the minimum and maximum available bandwidth values of the consolidated slice; and
iterating through each slice to assign the minimum and maximum available bandwidth values per slice to the minimum bandwidth available.

6. The method of claim 1 wherein the minimum and maximum available bandwidth is measured by a sampling process performed at defined measurement intervals on the order of one minute each.

7. The method of claim 6 wherein the minimum bandwidth available (min_bw_avail) is equal to the difference between a maximum system bandwidth and a maximum bandwidth used during the slice period divided by the maximum system bandwidth, and wherein the maximum bandwidth available (max_bw_avail) is equal to the difference between maximum system bandwidth and the minimum bandwidth used during the slice period divided by the maximum system bandwidth.

8. The method of claim 7 further comprising:
collecting network traffic information including network bandwidth, and a network uplink speed for the backup operation;
determining a size of a dataset to be saved in the backup operation;
determining an available bandwidth in each transfer window of a plurality of transfer windows for transfer data between the source and destination for the backup operation;
using the consolidated slices to define a transfer window of the plurality of transfer windows;
determining an order of the backup operations based on first ordering the backups based on decreasing dataset size and second ordering the transfer windows in order of decreasing bandwidth; and
matching the backups to the transfer windows in accordance with the first ordering and second ordering.

9. The method of claim 8 further comprising initiating a backup during a time window selected by the matching step, wherein the backup comprises a plurality of stages comprising taking a snapshot copy of a respective dataset from a source, transferring the dataset to a specified destination, tiering of the dataset to cloud storage, or replicating the dataset to secondary storage.

10. The method of claim 9 wherein the backup operations each implement a policy that specifies a plurality of service requirements and that is initiated at a time within the plurality of transfer windows.

11. The method of claim 10 wherein the available bandwidth is determined by monitoring network devices coupling the source and destination and comprising out-of-band devices coupled to other network equipment devices through a network implementing out-of-band protocols, and comprise at least one of managed switches, routers, or firewall devices, and further wherein the network comprises the out-of-band protocol network and a production software network coupling a plurality of computers together for the execution of one or more applications.

12. The method of claim 11 further comprising:
monitoring a flow of data between the source and destination for the respective backup;
compiling historical data of operational parameters for data transfers between the source and destination for the plurality of backups; and
determining an optimum time within the plurality of transfer windows based on the dataset size and the historical data for each backup of the plurality of backups.

13. A system for determining an optimal window size for backing up data in a series of time windows of a backup operation, comprising:
a first component dividing a backup window of the backup operation into a plurality of slices, assigning a respective percentage of a minimum and maximum available bandwidth to each slice of the plurality of slices, and iteratively determining which slices that have more than a certain percent difference between the assigned minimum and maximum available bandwidth;
a second component splitting the determined slices into sub-slices into 1-minute, and assigning the minimum and maximum available bandwidth to the sub-slices, determining which immediately consecutive slices have, when combined, a bandwidth difference that is less than or equal to the certain percent difference to identify consecutive under-utilized slices, and consolidating the consecutive under-utilized slices together.

14. The system of claim 13 further comprising an additional component processing all remaining sub-slices for consolidation into consolidated slices, and assigning the minimum and maximum available bandwidth values to each consolidated slice.

15. The system of claim 14 wherein at least one component computes an overall bandwidth loss, wherein the overall bandwidth loss is an average of minimum and maximum bandwidth losses multiplied by the percentage of time each slice occupies in the entire backup window.

16. The system of claim 15 further comprising a cleanup component determining if the total loss is less than a value that represents an optimal value defined by one or more threshold conditions comprising: a total loss is less than a first defined threshold, or an improvement in bandwidth loss less than a second defined threshold value from a prior iteration, or if X is less than a third minimum threshold value, and proceeding to a cleanup process if the total loss is less than the value, or iterating to a next time slice if the total loss is not less than the value.

17. The system of claim 16 wherein the cleanup process compris

Assignees

• Dell Products Lp

Inventors

• Malamut Mark
• Brenner Adam
• Murti Arun

Classifications

• G06F11/1464

  for networked environments · CPC title

• G06F11/1448 Primary

  Management of the data involved in backup or backup restore · CPC title

• G06F11/1461 Primary

  Backup scheduling policy · CPC title