# US 2022/0286501 A1 · Systems and methods for rate-based load balancing

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

- Number: US 2022/0286501 A1
- Document ID: US20220286501A1
- Status: Published application
- Kind: A1
- Application number: 17/751,493
- Filed: 2022-05-23
- Published: 2022-09-08
- Earliest claimed: 2019-08-21
- Assignees: Meta Platforms, Inc.
- Applicants: Meta Platforms, Inc.
- Inventors: Gian-Paolo MUSUMECI; Cong CHEN; Kestutis PATIEJUNAS
- Web page: https://patentgrep.com/patent/US20220286501A1

## CPC

- H04L 67/1008: based on parameters of servers, e.g. available memory or workload (monitoring of computer activity G06F11/30)
- G06N 3/0895: Weakly supervised learning, e.g. semi-supervised or self-supervised learning
- G06N 3/09: Supervised learning
- H04L 41/16: using machine learning or artificial intelligence
- H04L 47/215: using token-bucket
- H04L 67/101: based on network conditions
- H04L 67/62: Establishing a time schedule for servicing the requests
- G06N 3/08: Learning methods
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06N 5/02: Knowledge representation; Symbolic representation
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

According to examples, a system for rate-based load balancing may include a processor and a memory storing instructions. The processor may, through execution of the instructions, cause the system to receive a request for processing. The system may further identify a target server to transmit the request using a rate-based load balancing technique. In some examples, the rate-based load balancing technique may include: selecting a server, from a plurality of servers, as a potential target; receiving a readiness indicator for the selected server; and designating the selected server as the target server based on the readiness indicator. The system may transmit the request to the target server for processing.

## Claims

### Claim at position 1

1-20. (canceled)

### Claim 21

A system, comprising:
a processor; and
a memory storing instructions, which when executed by the processor, cause the processor to:
select a server, from a plurality of servers, as a potential target;
receive a readiness indicator for the selected server, wherein the readiness indicator is based on a timing technique using a fixed-capacity token bucket and a classification algorithm to determine if the selected server is to be designated as a target server; and
designate the selected server as the target server based on the readiness indicator; and
transmit a request to the target server for processing.

### Claim 22 (depends on 21)

The system of claim 21, wherein the timing technique measures and identifies whether a predetermined time period has passed.

### Claim 23 (depends on 22)

The system of claim 22, wherein a rate controller associated with the selected server generates the readiness indicator based on the timing technique, wherein the readiness indicator is a digital transmission that indicates whether the selected server is able to process the request.

### Claim 24 (depends on 23)

The system of claim 23, wherein the rate controller provides the readiness indicator upon selection of the server, by the processor, as the potential target or following passage of the predetermined time period.

### Claim 25 (depends on 24)

The system of claim 24, wherein the predetermined time period is configured by a user.

### Claim 26 (depends on 24)

The system of claim 24, wherein the predetermined time period is configured based on data associated with at least the request or the plurality of servers.

### Claim 27 (depends on 24)

The system of claim 24, wherein the predetermined time period is configured by using an artificial intelligence (AI) based machine learning technique based on information associated with the request or the selected server, wherein the artificial intelligence (AI) based machine learning technique utilizes at least one of a neural network, a tree-based model, a Bayesian network, a support vector, clustering, a kernel method, a spline, or a knowledge graph.

### Claim 28 (depends on 21)

The system of claim 21, wherein the classification algorithm provides assignment of instances to pre-defined classes to decide whether there are matches or correlation and into which one or more tokens, each comprising a packet of predetermined size, are added into the fixed-capacity token bucket at a predetermined fixed rate of lir seconds, where r is an integer, such that, in the event a token arrives when the fixed-capacity token bucket is full, the token is discarded and the fixed-capacity token bucket remains full until the selected server is designated as the target server based on the readiness indicator.

### Claim 29 (depends on 21)

The system of claim 21, wherein the system is an online system for processing digital requests associated with at least one of advertisements, payments, or social media.

### Claim 30

A method, comprising:
selecting a server, from a plurality of servers, as a potential target;
receiving a readiness indicator for the selected server, wherein the readiness indicator is based on a timing technique using a fixed-capacity token bucket and a classification algorithm to determine if the selected server is to be designated as a target server; and
designating the selected server as the target server based on the readiness indicator; and
transmitting a request to the target server for processing.

### Claim 31 (depends on 30)

The method of claim 30, wherein the timing technique measures and identifies whether a predetermined time period has passed.

### Claim 32 (depends on 31)

The method of claim 31, wherein a rate controller associated with the selected resource generates the readiness indicator based on the timing technique, wherein the readiness indicator is a digital transmission that indicates whether the selected resource is able to process the request.

### Claim 33 (depends on 32)

The method of claim 32, wherein the rate controller provides the readiness indicator upon selection of the resource, by the processor, as the potential target or following termination of a predetermined time period.

### Claim 34 (depends on 33)

The method of claim 33, wherein the predetermined time period is configured by a user.

### Claim 35 (depends on 33)

The method of claim 33, wherein the predetermined time period is configured based on data associated with at least the request or the plurality of resources.

### Claim 36 (depends on 33)

The method of claim 33, wherein the predetermined time period is configured by using an artificial intelligence (AI) based machine learning technique based on information associated with the request or the selected resource, wherein the artificial intelligence (AI) based machine learning technique utilizes at least one of a neural network, a tree-based model, a Bayesian network, a support vector, clustering, a kernel method, a spline, or a knowledge graph.

### Claim 37 (depends on 30)

The method of claim 30, wherein the classification algorithm provides assignment of instances to pre-defined classes to decide whether there are matches or correlation and into which one or more tokens, each comprising a packet of predetermined size, are added into the fixed-capacity token bucket at a predetermined fixed rate of 1/r seconds, where r is an integer, such that, in the event a token arrives when the fixed-capacity token bucket is full, the token is discarded and the fixed-capacity token bucket remains full until the selected server is designated as the target server based on the readiness indicator.

### Claim 38 (depends on 30)

The method of claim 30, wherein the rate-based load balancing system is an online system for processing digital requests associated with at least one of advertisements, payments, or social media.

### Claim 39

A non-transitory computer-readable storage medium having an executable stored thereon, which when executed instructs a processor to:
selecting a server, from a plurality of servers, as a potential target;
receiving a readiness indicator for the selected server, wherein the readiness indicator is based on a timing technique using a fixed-capacity token bucket and a classification algorithm to determine if the selected server is to be designated as a target server; and
designating the selected server as the target server based on the readiness indicator; and
transmitting a request to the target server for processing.

### Claim 40 (depends on 39)

The non-transitory computer-readable storage medium of claim 39, wherein the readiness indicator is provided by a rate controller associated with the selected resource following passage of a predetermined time period.