# US 2019/0379610 A1 · On-demand packet queuing in a network device

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

- Number: US 2019/0379610 A1
- Document ID: US20190379610A1
- Status: Published application
- Kind: A1
- Application number: 16/550,007
- Filed: 2019-08-23
- Published: 2019-12-12
- Earliest claimed: 2019-06-28
- Assignees: Intel Corporation
- Applicants: Intel Corporation
- Inventors: Arvind SRINIVASAN; Robert SOUTHWORTH; Helia A. NAEIMI
- Web page: https://patentgrep.com/patent/US20190379610A1

## CPC

- H04L 47/6255: {queue load conditions, e.g. longest queue first}
- H04L 47/11: Identifying congestion
- H04L 47/32: by discarding or delaying data units, e.g. packets or frames
- H04L 47/627: {policing}
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

Examples herein relate to allocation of an intermediate queue to a flow or traffic class (or other allocation) of packets prior to transmission to a network. Various types of intermediate queues are available for selection. An intermediate queue can be shallow and have an associated throughput that attempts to meet or exceed latency guarantees for a packet flow or traffic class. Another intermediate queue is larger in size and expandable and can be used for packets that are sensitive to egress port incast such as latency sensitive packets. Yet another intermediate queue is expandable but provides no guarantee on maximum end-to-end latency and can be used for packets where dropping is to be avoided. Intermediate queues can be deallocated after a flow or traffic class ends and related memory space can be used for another intermediate queue.

## Claims

### Claim 1

A network device comprising:
an ingress port to receive packets;
a processor to determine a flow associated with received packets; and
a queue assigner to determine an intermediate queue from a set of intermediate queues to assign to the received packets based on a policy, wherein the set of intermediate queues comprises one or more of:
a first type of intermediate queue that has an associated average throughput level or
a second type of intermediate queue that is deeper than the first type of intermediate queue and is expandable in size.

### Claim 2 (depends on 1)

The network device of claim 1, wherein the set of intermediate queues comprises a third type of intermediate queue that is expandable in size but shallower than the second type of intermediate queue.

### Claim 3 (depends on 1)

The network device of claim 1, wherein the policy comprises an allocation of an intermediate queue based on one or more of: egress port congestion, egress port incast, or flow life length.

### Claim 4 (depends on 1)

The network device of claim 1, comprising:
an egress port;
a scheduler; and
at least one egress queue associated with the egress port, the at least one egress queue to receive packets from one or more intermediate queues and the scheduler to schedule transmission of packets from an egress queue.

### Claim 5 (depends on 1)

The network device of claim 1, wherein the first type of intermediate queue comprises an intermediate queue with a guaranteed maximum ingress-to-egress port latency time and packets can be dropped from the first type of intermediate queue due to overflow.

### Claim 6 (depends on 1)

The network device of claim 1, wherein the first type of intermediate queue comprises an egress queue for one or more egress ports.

### Claim 7 (depends on 1)

The network device of claim 1, wherein the second type of intermediate queue is expanded based on packet drop from the second type of intermediate queue.

### Claim 8 (depends on 1)

The network device of claim 1, wherein the second type of intermediate queue is associated with no guarantee on maximum end-to-end latency.

### Claim 9 (depends on 2)

The network device of claim 2, wherein the third type of intermediate queue is expanded based on egress port congestion or incast detection.

### Claim 10 (depends on 1)

The network device of claim 1, wherein the intermediate queues are allocated in a memory device.

### Claim 11 (depends on 1)

The network device of claim 1, wherein the queue assigner is to deallocate an intermediate queue allocated to a flow based on changes in one or more of: egress port congestion, egress port incast, or flow life length.

### Claim 12 (depends on 1)

The network device of claim 1, comprising one or more of: a server, data center, rack, network interface, or host computer.

### Claim 13

A method performed in a network device comprising:
receiving a request for an intermediate queue and
allocating an intermediate queue from a set of intermediate queues based on an applicable policy, the set of intermediate queues comprising at least:
a first type of intermediate queue that is not expandable and has an associated throughput level,
a second type of intermediate queue that is expandable, or
a third type of intermediate queue that is deeper than the second type of intermediate queue and is expandable.

### Claim 14 (depends on 13)

The method of claim 13, wherein the policy comprises an allocation of an intermediate queue based on one or more of: egress port congestion, egress port incast, or flow life length.

### Claim 15 (depends on 13)

The method of claim 13, wherein:
the first type of intermediate queue comprises a shallow intermediate queue with a guaranteed maximum ingress-to-egress port latency time and permits dropping of packets due to overflow,
the second type of intermediate queue is expandable in an event of a congestion event or incast detection, and
the third type of intermediate queue is expanded if a packet drop occurs and the third type of intermediate queue provides no guarantee on maximum end-to-end latency.

### Claim 16 (depends on 13)

The method of claim 13, comprising deallocating an intermediate queue based on end of a flow.

### Claim 17 (depends on 13)

The method of claim 13, wherein allocating an intermediate queue from a set of intermediate queues based on an applicable policy comprises considering telemetry data, the telemetry data comprising congestion level at an egress port, bandwidth utilization at an ingress port, bandwidth utilization at egress port, or bandwidth utilized by a specific flow.

### Claim 18

A system comprising:
one or more processors;
a memory; and
a network device to receive a packet to be egressed to a network medium, the network device to determine an intermediate queue to allocate to the packet, the network device to
allocate an intermediate queue in the memory to the packet based on one or more of: level or permitted latency of receipt to transmit, sensitivity to packet drop, or flow life length and
assign the packet to the allocated intermediate queue.

### Claim 19 (depends on 18)

The system of claim 18, wherein based on an end of a flow associated with the packet, the network device is to deallocate an intermediate queue and allow any portion of the deallocated queue to be used for allocation of another intermediate queue.

### Claim 20 (depends on 18)

The system of claim 18, wherein:
a first type of intermediate queue comprises a shallow intermediate queue with a guaranteed maximum ingress-to-egress port latency time and permits dropping of packets due to overflow,
a second type of intermediate queue is expandable based on a congestion event or incast detection, and
a third type of intermediate queue is expanded if a packet drop occurs and the third type of intermediate queue provides no guarantee on maximum end-to-end latency.

### Claim 21 (depends on 18)

The system of claim 18, wherein the intermediate queues are allocated in the memory.
