# US 2026/0089027 A1 · Systems, apparatuses and methods for managing and routing data in a multi-tier network architecture

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

- Number: US 2026/0089027 A1
- Document ID: US20260089027A1
- Status: Published application
- Kind: A1
- Application number: 18/893,112
- Filed: 2024-09-23
- Published: 2026-03-26
- Earliest claimed: 2024-09-23
- Applicants: Huawei Technologies Co., Ltd.
- Inventors: Amr Adel Nasr Abdelnasser; Xingjun Chu; Ashkan Sobhani
- Web page: https://patentgrep.com/patent/US20260089027A1

## CPC

- H04L 12/44: Star or tree networks
- H04L 45/025: {Updating only a limited number of routers, e.g. fish-eye update}
- H04L 45/08: {Learning-based routing, e.g. using neural networks or artificial intelligence}
- H04L 45/24: Multipath
- H04L 45/48: Routing tree calculation
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

Methods, systems, and apparatuses are provided for managing and routing data in a multi-tier network architecture. The methods include transmitting a data flow comprising a sequence of data packets to a source leaf switch and determining if an intended route exists in a forwarding table stored at the source leaf switch using routing metadata. If the intended route does not exist, a control packet is transmitted to a centralized coordinator switch, which modifies an allocation table to include an updated route. The forwarding table at the source leaf switch is then updated with the updated route, and the data flow is transmitted to its destination.

## Claims

### Claim 1

A method for routing data in a multi-tier network architecture, comprising:
transmitting a data flow to a source leaf switch of a plurality of leaf switches in the multi-tier network architecture, the data flow comprising a sequence of data packets;
determining if an intended route for the data flow exists in a forwarding table stored at the source leaf switch using routing metadata from at least one of the data packets;
in response to determining that the intended route does not exist in the forwarding table, transmitting a control packet to a centralized coordinator switch;
modifying an allocation table stored at the centralized coordinator switch to include an updated route for the data flow based on current network conditions;
updating the forwarding table stored at the source leaf switch using the updated route; and
transmitting the data flow to a destination node according to the updated forwarding table.

### Claim 2 (depends on 1)

The method of claim 1, wherein modifying the allocation table comprises utilizing a loop algorithm to determine the updated route for the data flow.

### Claim 3 (depends on 1)

The method of claim 1, further comprising:
classifying the data flow as a mice flow or an elephant flow based on a size of the data flow; and
in response to the data flow being classified as the mice flow, routing the data flow according to an Equal-Cost Multi-Path (ECMP) or Hop-by-hop Utilization-aware Load balancing Architecture (HULA) scheme.

### Claim 4 (depends on 1)

The method of claim 1, wherein the multi-tier network architecture comprises a first tier and a second tier, the first tier comprising the plurality of leaf switches and the second tier comprising a plurality of spine switches, the centralized coordinator switch being one of the plurality of spine switches.

### Claim 5 (depends on 1)

The method of claim 1, wherein the multi-tier network architecture comprises a plurality of layers to allow simultaneous data transmission via a same route on corresponding layers, and
wherein the allocation table further comprises a sending layer table and a receiving layer table for determining which of the plurality of layers the data flow is to use.

### Claim 6 (depends on 1)

The method of claim 1, wherein the multi-tier network architecture comprises a first tier comprising the plurality of leaf switches, a second tier comprising a plurality of spine switches, and a third tier comprising a plurality of core switches, the centralized coordinator switch being one of the plurality of core switches.

### Claim 7 (depends on 6)

The method of claim 6, wherein:
a first number of the spine switches and a first number of the core switches are grouped as a first group, and a second number of the spine switches and a second number of the core switches are grouped as a second group, and
the control packet comprises a first control packet and a second control packet, the first control packet being transmitted to the centralized coordinator switch acting as a first-stage coordinator for determining whether the first group or the second group is included in the updated route, the second control packet being transmitted to one of the core switches in the determined first or second group acting as a second-stage coordinator for determining which of the plurality of core switches is included in the updated route.

### Claim 8 (depends on 6)

The method of claim 6, wherein:
the plurality of leaf switches and the plurality of spine switches are grouped as a plurality of pods, each comprising at least two of the plurality of leaf switches and at least two of the plurality of spine switches, and
the control packet is transmitted to the centralized coordinator switch for determining which of the plurality of pods are included in the updated route.

### Claim 9 (depends on 1)

The method of claim 1, wherein:
the multi-tier network architecture comprises a first tier and a second tier, the first tier comprising the plurality of leaf switches and the second tier comprising a plurality of spine switches,
the multi-tier network architecture comprises a plurality of physical links between each of the plurality of the leaf switches and each of the plurality of spine switches to allow simultaneous data transmission via a same route on corresponding physical links, and
the allocation table further comprises a link index for determining which of the plurality of physical links the data flow is to use.

### Claim 10 (depends on 1)

The method of claim 1, further comprising:
in response to completing the transmission of the data flow,
retaining the forwarding table for a predetermined period; and
erasing information on the data flow from the allocation table to free up resources for future routing tasks.

### Claim 11

A network system comprising:
a plurality of leaf switches and a centralized coordinator switch, the leaf switches and the centralized coordinator switch being part of a multi-tier network architecture;
wherein a source leaf switch of the plurality of leaf switches comprises:
a packet receiver for receiving a data flow comprising a sequence of data packets;
a leaf storage unit, in which a forwarding table is stored; and
a leaf control unit adapted to determine if an intended route for the data flow exists in the forwarding table using routing metadata from at least one of the data packets, and, in response to determining that the intended route does not exist in the forwarding table, to generate a control packet that is to be transmitted to the centralized coordinator switch;
wherein the centralized coordinator switch comprises:
a coordinator storage unit, in which an allocation table is stored; and
a coordinator control unit adapted to modify the allocation table to include an updated route for the data flow based on current network conditions, and to transmit the updated route to the source leaf switch; and
wherein the leaf control unit of the source leaf switch is further adapted to update the forwarding table using the updated route, and to transmit the data flow to a destination node according to the updated forwarding table.

### Claim 12 (depends on 11)

The network system of claim 11, wherein the allocation table is modified by utilizing a loop algorithm to determine the updated route for the data flow.

### Claim 13 (depends on 11)

The network system of claim 11, wherein the source leaf switch further comprises a classification unit configured to classify the data flow as a mice flow or an elephant flow based on a size of the data flow; and
wherein the data flow is routed according to an Equal-Cost Multi-Path (ECMP) or Hop-by-hop Utilization-aware Load balancing Architecture (HULA) scheme in response to the data flow being classified as the mice flow.

### Claim 14 (depends on 11)

The network system of claim 11, wherein the multi-tier network architecture comprises:
a first tier comprising the plurality of leaf switches; and
a second tier comprising a plurality of spine switches, wherein the centralized coordinator switch is one of the plurality of spine switches.

### Claim 15 (depends on 11)

The network system of claim 11, wherein the multi-tier network architecture comprises:
a plurality of layers for allowing simultaneous data transmission via a same route on corresponding layers,
wherein the allocation table further comprises a sending layer table and a receiving layer table for determining which of the plurality of layers the data flow is to use.

### Claim 16 (depends on 11)

The network system of claim 11, wherein the multi-tier network architecture comprises:
a first tier comprising the plurality of leaf switches;
a second tier comprising a plurality of spine switches; and
a third tier comprising a plurality of core switches,
wherein the centralized coordinator switch is one of the plurality of core switches.

### Claim 17 (depends on 16)

The network system of claim 16, wherein:
a first number of the spine switches and a first number of the core switches are grouped as a first group, and a second number of the spine switches and a second number of the core switches are grouped as a second group, and
the control packet comprises a first control packet and a second control packet, the first control packet being transmitted to the centralized coordinator switch acting as a first-stage coordinator for determining whether the first group or the second group is included in the updated route, the second control packet being transmitted to one of the core switches in the determined first or second group acting as a second-stage coordinator for determining which of the plurality of core switches is included in the updated route.

### Claim 18 (depends on 16)

The network system of claim 16, wherein:
the plurality of leaf switches and the plurality of spine switches are grouped as a plurality of pods, each pod comprising at least two of the plurality of leaf switches and at least two of the plurality of spine switches, and
the control packet is transmitted to the centralized coordinator switch for determining which of the plurality of pods are included in the updated route.

### Claim 19 (depends on 11)

The network system of claim 11, wherein:
the multi-tier network architecture comprises a first tier and a second tier, the first tier comprising the plurality of leaf switches and the second tier comprising a plurality of spine switches,
the multi-tier network architecture comprises a plurality of physical links between each of the plurality of the leaf switches and each of the plurality of spine switches to allow simultaneous data transmission via a same route on corresponding physical links, and
the allocation table further comprises a link index for determining which of the plurality of physical links the data flow is to use.

### Claim 20

An apparatus comprising:
a coordinator storage unit, in which an allocation table is stored; and
a coordinator control unit adapted to:
receive a control packet from a source leaf switch of a plurality of leaf switches in a multi-tier network architecture, the control packet being generated in response to the source leaf switch determining that an intended route for a data flow comprising a sequence of data packets does not exist in a forwarding table stored at the source leaf switch using routing metadata from at least one of the data packets;
modify the allocation table to include an updated route for the data flow based on current network conditions; and
transmit the updated route to the source leaf switch.
