# US 2025/0156111 A1 · Flexible Write Cache Policy in RAID Environments

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0156111 A1
- Document ID: US20250156111A1
- Status: Published application
- Kind: A1
- Application number: 18/507,005
- Filed: 2023-11-10
- Published: 2025-05-15
- Earliest claimed: 2023-11-10
- Assignees: Avago Technologies International Sales Pte. Limited
- Applicants: Avago Technologies International Sales Pte. Limited
- Inventors: Sourav Saha
- Web page: https://patentgrep.com/patent/US20250156111A1

## CPC

- G06F 3/0659: {Command handling arrangements, e.g. command buffers, queues, command scheduling}
- G06F 12/0868: Data transfer between cache memory and other subsystems, e.g. storage devices or host systems
- G06F 3/0604: {Improving or facilitating administration, e.g. storage management}
- G06F 3/0664: {at device level, e.g. emulation of a storage device or system}
- G06F 3/0689: {Disk arrays, e.g. RAID, JBOD}
- G06F 12/0804: with main memory updating (G06F12/0806 takes precedence)
- G06F 12/0871: Allocation or management of cache space
- G06F 2212/1016: Performance improvement
- G06F 2212/222: Non-volatile memory
- G06F 2212/262: configured as RAID
- G06F 2212/312: In storage controller
- G06F 2212/502: using adaptive policy
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Solutions for providing more customizable performance for a virtual disk in a RAID environment. Some solutions provide for a flexible write cache policy that manages a proportion of input-output operations written in write-back mode against input-output operations written in write-though mode. In some cases, aspects of the flexible write cache policy can be specified by a user.

## Claims

### Claim 1

A redundant array of independent disks (RAID) controller to manage a virtual disk comprising a plurality of arms, each of the arms comprising physical storage on a physical disk, the RAID controller comprising:
a cache comprising a random-access memory (RAM);
a virtual disk property table (VDPT);
a cache manager comprising:
hardware circuitry to receive a first plurality of write input-output operations (IO);
hardware circuitry to execute each of the first plurality of IOs in write-back mode; and
hardware circuitry to transmit a completion confirmation for each completed IO executed in write-back mode;
an IO manager comprising:
hardware circuitry to receive a second plurality of IOs;
hardware circuitry to execute each of the second plurality of IOs on the virtual disk; and
hardware circuitry to transmit a completion confirmation for each IO executed on the virtual disk;
firmware comprising:
logic to receive user input indicating a preferred ratio of write-back IOs to total IOs;
logic to establish a flexible write cache policy based at least on the preferred ratio of write-back IOs to total IOs, the flexible write cache policy specifying:
a first proportion of IOs to be performed in write-back mode (PC); and
a second proportion of IOs to be performed in write-though mode (PD);
logic to store the flexible write cache policy in the VDPT;
logic to set a counter to implement the flexible write cache policy, wherein the counter implements the flexible write cache policy using the following relationships:
W
tot
=
(
W
C
+
W
D
)
P
C
=
W
C
W
tot
P
D
=
W
D
W
tot
=
W
tot
-
W
C
W
tot
where Wtot represents a total number of IOs executed per cycle as specified by an IO group size of the flexible write cache policy, WC represents a number of IOs executed in write-back mode per cycle, and WD represents a number of IOs executed in write-through mode per cycle; and
a command dispatcher (CDU) comprising:
hardware circuitry to store the counter;
hardware circuitry to receive a plurality of host IOs;
hardware circuitry to perform a repeating cycle comprising:
(1) creating of an accelerated IO (ACIO) from each of the host IOs
(2) transmitting each remaining ACIO to the cache manager for execution until the counter exceeds WC;
(3) updating the counter for every ACIO transmitted to the cache manager;
(4) transmitting each remaining ACIO to the IO manager for execution until the counter reaches Wtot;;
(5) updating the counter for every ACIO transmitted to the IO manager; and
(6) resetting the counter after transmitting Wtot ACIOs;
hardware circuitry to receive a plurality of write completion confirmations from the cache manager and the IO manager; and
hardware circuitry to notify the host of each host IO completed.

### Claim 2

A device, comprising
logic to manage a virtual disk comprising a plurality of arms, each of the arms comprising physical storage on a physical disk;
a cache comprising random access memory (RAM);
logic to establish a flexible write cache policy;
logic to execute a first plurality of write input-output operations (IO) in write-back mode, based at least in part on the flexible write cache policy; and
logic to execute a second plurality of IOs in write-though mode, based at least in part on the flexible write cache policy.

### Claim 3 (depends on 2)

The device of claim 2, further comprising:
logic to receive a plurality of IOs;
wherein:
the logic to execute a first plurality of write input-output operations (IO) in write-back mode comprises:
logic to transmit a first set of the plurality of IOs for execution in write-back mode; and
the logic to execute a second plurality of IOs in write-though mode comprises:
logic to transmit a second set of the plurality of IOs for execution on the virtual disk.

### Claim 4 (depends on 2)

The device of claim 2, further comprising:
logic to set a counter to implement the flexible write cache policy.

### Claim 5 (depends on 4)

The device of claim 4, further comprising:
logic to receive user input indicating a preferred ratio of write-back IOs to total IOs (PC));
wherein the flexible write cache policy implements the preferred ratio (PC).

### Claim 6 (depends on 5)

The device of claim 5, wherein the preferred ratio (PC) of write-back IOs to total IOs is expressed as one of a limited plurality of discrete percentages or fractions.

### Claim 7 (depends on 5)

The device of claim 5, wherein:
the flexible write cache policy establishes a group size Wtot;
the device further comprises:
logic to configure the counter to reset after Wtot IOs have been performed;
the logic to execute a first plurality of IOs in write-back mode comprises:
logic to perform (PC·Wtot) IOs in write-back mode before the counter resets; and
the logic to execute a second plurality of IOs in write-though mode comprises:
logic to perform ((1−PC)·Wtot) IOs in write-though mode before the counter resets.

### Claim 8 (depends on 7)

The device of claim 7, wherein Wtot is 5 and (PC·Wtot) is an integer between 0 and 5.

### Claim 9 (depends on 6)

The device of claim 6, wherein the logic to set the counter comprises logic to reset the counter at start of day (SOD).

### Claim 10 (depends on 2)

The device of claim 2, further comprising:
a processor;
hardware circuitry; and
a firmware that stores instructions executable by the processor, wherein:
the logic to establish the flexible write cache policy comprises:
instructions stored in the firmware executable by the processor to establish the flexible write cache policy.

### Claim 11 (depends on 10)

The device of claim 10, wherein the logic comprises the hardware circuitry.

### Claim 12 (depends on 10)

The device of claim 10, wherein:
the hardware circuitry stores a virtual disk property table (VDPT);
the instructions to establish the flexible write cache policy comprise:
instructions executable by the processor to store the flexible write cache policy in the VDPT.

### Claim 13 (depends on 10)

The device of claim 10, wherein:
the first plurality of IOs comprises a first plurality of host IOs;
the second plurality of IOs comprises a second plurality of host IOs;
the logic to execute the first plurality of IOs in write-back mode comprises:
hardware circuitry to create a first plurality of accelerated IOs (ACIO) from the first plurality of host IOs; and
the logic to execute the second plurality of IOs in write-though mode comprises:
hardware circuitry to create a second plurality of ACIOs from the second plurality of host IOs.

### Claim 14 (depends on 13)

The device of claim 13, wherein the device further comprises:
hardware circuitry to receive a notification of each completed ACIO; and
hardware circuitry to notify the host of each completed host IO.

### Claim 15 (depends on 13)

The device of claim 13, wherein:
the ACIOs are parent ACIOs; and
the device further comprises:
hardware circuitry to create a set of two or more child ACIOs from at least one of the parent ACIOs.

### Claim 16 (depends on 15)

The device of claim 15, wherein:
the logic to execute the first plurality of IOs in write-back mode further comprises hardware circuitry to direct the first plurality of ACIOs to a first hardware component of the device; and
the logic to execute the second plurality of IOs in write-though mode further comprises hardware circuitry to direct the second plurality of ACIOs to a second hardware component of the device.

### Claim 17 (depends on 16)

The device of claim 16, wherein the logic to execute the first plurality of IOs in write-back mode further comprises:
hardware circuitry in the first component to receive the first plurality of ACIOs; and
hardware circuitry in the first component to write the first plurality of ACIOs to the cache; and
hardware circuitry in the first component to transmit a completion confirmation after each of the ACIOs has been written to the cache.

### Claim 18 (depends on 17)

The device of claim 17, wherein the logic to execute the second plurality of IOs on the virtual disk comprises:
hardware circuitry in the second component to receive a plurality of ACIOs;
hardware circuitry in the second component to create one or more drive IOs for each of the received plurality of ACIOs;
hardware circuitry in the second component to send each of the drive IOs to the virtual disk for execution; and
hardware circuitry in the second component to transmit a completion confirmation after the one or more drive IOs for each of the received ACIOs have been executed on the virtual disk.

### Claim 19 (depends on 17)

The device of claim 17, wherein:
the hardware circuitry to direct the second plurality of ACIOs to a second hardware component of the device comprises hardware circuitry to transmit the second plurality of ACIOs to a third component;
the third component comprises:
hardware circuitry to generate one or more parity values;
hardware circuitry to write the second plurality of ACIOs and the one or more parity values to the cache;
hardware circuitry to generate a third set of ACIOs from the second plurality of ACIOs and the one or more parity values; and
hardware circuitry to flush the cache by directing the third set of ACIOs to the second component; and
the plurality of ACIOs received by the second component comprises the third set of ACIOs.

### Claim 20

A method, comprising:
managing a virtual disk comprising a plurality of arms, each of the arms comprising physical storage on a physical disk;
establishing a flexible write cache policy;
executing a first plurality of write input-output operations (IO) on a cache, based at least in part on the flexible write cache policy; and
executing a second plurality of IOs in write-though mode, based at least in part on the flexible write cache policy.
