Login method based on edge side AI computing power device SN

Abstract

The application relates to the technical field of Internet of Things, and discloses a login method based on an edge side AI computing power equipment SN, which comprises the following steps: an AI edge computing power gateway periodically performs a network detection task, and determines a network state through double-layer connectivity verification; when it is determined that the network is disconnected, the AI edge computing power gateway starts an AP temporary hotspot to perform display-free remote network configuration of a network configuration terminal; when network configuration is completed or the network is recovered, the AI edge computing power gateway is switched to a STA mode and differentiates and reports a device solidification SN and a local area network IP of the AI edge computing power gateway to an AIDA management backend; an AIDA client queries mapping data of the device solidification SN and the local area network IP in the AIDA management backend based on the device solidification SN, and completes IP-free login. By using the application scheme, automatic hotspot network configuration, intelligent network state switching, SN binding IP automatic reporting and client SN input IP-free login of the computing power gateway can be realized.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) communication technology, and more specifically to a login method based on an edge-side AI computing power device (SN). Background Technology

[0002] With the rapid popularization of local large-scale models and edge AI intelligent agents, edge AI computing power gateways are being deployed on a large scale in various private LAN environments such as offices and homes, forming a localized private AI computing power service system. These computing power gateway devices deploy large-scale models locally and are equipped with AI intelligent agents capable of autonomously scheduling tasks. They support users in completing personalized workflow customization on demand. All inference calculations and business data are processed in a closed loop within the LAN, without relying on external cloud computing power, effectively ensuring the security of users' private data. At the same time, multiple access roles often exist within the same LAN, with different personnel possessing different device operation permissions. This necessitates hierarchical login management and the isolation of each person's AI business data, placing higher demands on the ease of access and the security of identity verification for the computing power gateway.

[0003] Currently, AI edge computing gateways in office and home private scenarios generally use DHCP (Dynamic Host Configuration Protocol) to dynamically assign IP addresses. These IP addresses are highly susceptible to change after device restarts, network reconnections, or the addition of new computing nodes. Existing operation and maintenance access solutions have several technical shortcomings that make them difficult to adapt to private intranet AI computing scenarios. First, traditional login methods require manual searching and input of device IP addresses. When deploying computing gateways in batches in office server rooms and multiple home nodes, the workload of manually registering and maintaining the IP list is enormous. After the IP is dynamically updated, the record table cannot be updated synchronously, which directly causes maintenance personnel to be unable to log in to the target device, interrupting local large model debugging, AI agent task configuration and other work. Secondly, private LANs have multi-role access requirements, and the general IP login mode lacks a hardware unique identity binding verification mechanism. It relies solely on IP to distinguish devices, which can easily lead to device confusion and unauthorized access by unauthorized personnel. This poses a risk of leakage of local private large models and user-customized workflow data. Furthermore, when multiple computing gateways are deployed in a distributed manner on the intranet, it is not possible to quickly scan device IPs in batches in cross-network segment and multi-subnet environments, making it difficult for operation and maintenance personnel to quickly locate the computing gateway corresponding to a specific intelligent agent or dedicated inference task. Finally, existing simple SN (Serial Number) and IP static mapping solutions are mostly geared towards general IoT devices in industrial and monitoring applications. They are not adapted to private local AI computing power scenarios, lack supporting logic for multi-role hierarchical login on the intranet, and lack the ability to automatically discover local devices when disconnected from the external network. Therefore, they cannot achieve automatic identification and addressing of computing power service devices based on SN in a pure intranet environment.

[0004] In summary, existing access and login solutions cannot meet the needs of convenient operation and maintenance and secure hierarchical access for large local models and AI intelligent agent clusters within office and home private LANs. There is an urgent need for a technical solution based on device hardware SN that can automatically discover computing power gateways on the intranet and log in without IP address. Summary of the Invention

[0005] This invention provides a login method based on the SN of an edge-side AI computing power device, which can realize automatic hotspot network configuration, intelligent network status switching, automatic reporting of SN-bound IP, and client login to the computing power gateway without IP by entering the SN.

[0006] Therefore, the present invention provides the following technical solution: A login method based on an edge-side AI computing power device SN, the method comprising: Step 1: The AI ​​edge computing gateway periodically performs network detection tasks and determines the network status through two-layer connectivity verification. Step 2: When the network is determined to be disconnected, the AI ​​edge computing gateway opens a temporary AP hotspot to perform remote network configuration of the network configuration terminal without a display. Step 3: When the network configuration is completed or the network is restored, the AI ​​edge computing power gateway switches to STA mode and reports the device's fixed SN and LAN IP to the AIDA management backend in a differentiated manner. Step 4: The AIDA client queries the mapping data between the device's fixed SN and the local area network IP in the AIDA management backend based on the device's fixed SN of the AI ​​edge computing gateway, and completes the IP-free login to the AI ​​edge computing gateway.

[0007] Optionally, step 1 includes: Step 11: The AI ​​edge computing gateway starts network detection 1 minute after powering on, and then performs a round of network verification every 1 minute thereafter. Step 12: The AI ​​edge computing gateway reads the device's fixed SN and reads the local area network IP; Step 13: The AI ​​edge computing gateway performs a two-layer connectivity check, which includes: a first layer of batch pinging of public DNS nodes; if any node is successfully pinged, the network is considered normal; a second layer of fallback interface check; if all ping probes fail, an attempt is made to access the network probe interface of the AIDA management backend to probe network connectivity; if both layers fail, the current network is considered abnormal, the network is disconnected, and the process proceeds to step 2.

[0008] Optionally, step 2 includes: Step 21: The AI ​​edge computing gateway sets the parameter information for the AP temporary hotspot; Step 22: The AI ​​edge computing gateway activates the AP temporary hotspot connection, and the user uses the network configuration terminal to scan and connect to the AP temporary hotspot; Step 23: After the network configuration terminal connects to the AP temporary hotspot, it accesses the network configuration page of the AI ​​edge computing power gateway, fills in the network configuration parameters such as the target local area network WiFi name and password for the office or home, and the AI ​​edge computing power gateway saves and starts the newly configured network.

[0009] Optionally, step 3 includes: Step 31: After the network configuration parameters are issued, the AI ​​edge computing power gateway shuts down the AP temporary hotspot, enables STA mode, and loads the WiFi credentials within the AI ​​edge computing power gateway to access the private local area network. Step 32: The AI ​​edge computing gateway performs a two-layer connectivity check again. If the external network is still unreachable, the AI ​​edge computing gateway automatically switches back to the AP temporary hotspot and continues to retain the temporary network distribution channel. Step 33: After the network is stably connected, the AI ​​edge computing power gateway reads the previously reported local area network IP cached in the AI ​​edge computing power gateway, compares it with the real-time local area network IP of the current distribution terminal, and reports the device fixed SN and local area network IP of the AI ​​edge computing power gateway to the AIDA management backend. Step 34: After receiving the data, the AIDA management backend stores the dynamic mapping relationship between the device's fixed SN and the local area network IP address of the AI ​​edge computing gateway using the device's fixed SN as the primary key. At the same time, it writes the expiration time to indicate the device's online validity period and constructs a dynamic SN-IP mapping database. Step 35: After the AI ​​edge computing gateway successfully reports the latest local area network IP address, write the latest IP address into the AI ​​edge computing gateway cache as the benchmark for the next round of comparison.

[0010] Optionally, in step 33, if the local area network IP remains unchanged, the AI ​​edge computing gateway does not perform a reporting operation; if the local area network IP changes, or if the cache of historical local area network IPs is not started for the first time, the AI ​​edge computing gateway reports the device's fixed SN and local area network IP to the AIDA management backend.

[0011] Optionally, step 3 further includes the AIDA management backend periodically scanning the dynamic SN-IP mapping database and clearing offline device records that have exceeded their validity period.

[0012] Optionally, step 4 includes: Step 41: The user opens the AIDA client and enters the device's fixed SN number as the target AI edge computing gateway; Step 42: The AIDA client first queries the historical login address of the AI ​​edge computing power gateway recorded locally; if the AIDA client cannot log in to the historical login address, it sends an address query request to the AIDA management backend carrying the device's fixed SN of the AI ​​edge computing power gateway. Step 43: The AIDA management backend retrieves the dynamic SN-IP mapping database, verifies the SN's legality and the device's online status, and returns the latest LAN IP bound to the SN after successful verification. Step 44: The AIDA client uses the latest LAN IP address obtained from the query to directly log in to the AI ​​edge computing gateway that is the login target; Step 45: After successful login, users can access the local large model, debug AI agents, and edit personalized workflows of the AI ​​computing power gateway according to their different role user permissions.

[0013] A login system based on the serial number (SN) of an edge-side AI computing device is disclosed. The system includes an AI edge computing gateway, an AIDA client, an AIDA management backend, and a network configuration terminal. The AI ​​edge computing gateway periodically performs network detection tasks and determines the network status through a two-layer connectivity check. When the network is determined to be disconnected, the AI ​​edge computing gateway activates a temporary AP hotspot to perform remote network configuration for the network configuration terminal without a monitor. When network configuration is completed or the network is restored, the AI ​​edge computing gateway switches to STA mode and reports the device's fixed SN and LAN IP address to the AIDA management backend in a differentiated manner. The AIDA client queries the mapping data between the device's fixed SN and LAN IP address in the AIDA management backend based on the device's fixed SN to complete the IP-free login.

[0014] A computer-readable storage medium, wherein the computer program, when executed by a processor, performs the steps of the login method based on an edge-side AI computing device SN.

[0015] This invention provides a login method based on the serial number (SN) of edge AI computing power devices, belonging to the fields of IoT communication, edge AI computing power device operation and maintenance, and remote access authentication. It achieves automatic activation of a wireless hotspot for remote network configuration without external devices in offline scenarios; it enables automatic reporting of hardware SN and dynamic mapping between the device and the local area network IP after the device is connected to the network; and it allows the client to automatically resolve the IP and discover internal network computing power service devices based on the SN serial number, providing a one-stop solution for the entire process of private edge computing power gateway deployment, addressing, and remote login. This invention solves the problems of cumbersome AI edge computing power gateway configuration, requiring external monitors, keyboards, and mice for local viewing and IP configuration, and high operational costs for batch deployment of office / home private computing power nodes; it also solves the problem of frequent DHCP dynamic IP changes of computing power gateways requiring manual on-site IP re-querying and ledger updates, leading to AI client login failures and interruptions in local large-scale model, AI agent debugging, and personalized workflow configuration; and it solves the problem of relying on manual tables to maintain the SN-IP mapping relationship, which is prone to input errors and ledger loss, requiring each device to be connected to an external monitor for troubleshooting after a lost record. This invention addresses the issues of low operational efficiency in P-type computing power devices. It resolves the problem of unauthorized access and data leakage risks associated with local private large-scale models and user-customized AI business data when relying solely on IP login without hardware serial number (SN) verification in multi-role user scenarios within a private LAN. It also addresses the time-consuming manual IP lookup process when deploying multiple computing power gateways in a distributed, cross-subnet environment within an intranet. Furthermore, it resolves the issue that users accessing computing power devices using AI clients must know the IP address beforehand, as automatic addressing and connection establishment based solely on device SN is insufficient. Compared to existing technologies, this invention offers the following technical advantages: First, this invention simplifies the batch configuration of private computing gateways without requiring external peripherals, eliminating the cumbersome IP address lookup process of traditional solutions. The `auto_hotspot.sh` script automatically activates AIDA-AP open hotspots when the network is disconnected. The hotspot uses a fixed network segment (10.43.0.1 / 24), allowing maintenance personnel to connect to the hotspot and distribute internal WiFi configurations using only tablets or laptops, eliminating the need for monitors, keyboards, and mice on local devices. Combined with a persistent virtual network interface (NIC) design (`wlan_sta`), it avoids network jitter caused by repeatedly creating and destroying NICs. Compared to existing technologies that require manual local connection checks of IP addresses and manual record-keeping, this method improves deployment efficiency by over 90% when deploying jointly issued computing gateways in batches for government and enterprise use, significantly reducing on-site implementation costs.

[0016] Secondly, the dual-layer network connectivity verification adapts to complex government and enterprise network environments, avoiding false alarms caused by single-probe methods. This invention employs a dual-layer detection logic of public DNS ping probing and official website network verification interface, setting parameters such as PING_COUNT=3, timeout of 4s, and a maximum interface request duration of 8s, making it compatible with special network scenarios in government and enterprise intranets that block ICMP packets. Existing technologies rely solely on a single ping to determine network status, which can easily lead to issues such as correctly identifying a network as offline or incorrectly starting a hotspot. This solution's dual-layer verification mechanism significantly reduces the probability of false network status judgments, ensuring the stable operation of network configuration and IP reporting processes.

[0017] Secondly, the differentiated IP reporting mechanism reduces invalid backend requests and lowers the load on the official website server. Devices cache historical IPs locally via `lastip.conf`, and only call the ` / api / v1 / device / ip / report` interface to report the serial number (SN) and local area network IP to `imaida.cn` when the device first comes online or when the IP changes. The scheduled detection period is one minute. Compared to existing technologies that continuously report all IPs and use high-frequency packet sending for port scanning, this significantly reduces network transmission and backend database read / write pressure. For government and enterprise scenarios deploying dozens of joint computing gateways in batches, server resource consumption is significantly reduced.

[0018] Then, based on globally unified addressing using hardware serial numbers (SNs), port scanning is abandoned, eliminating issues of internal network bandwidth consumption and device confusion. Utilizing the official website's backend to store dynamic mapping relationships with the SN as the primary key, users can obtain an IP address simply by entering the device serial number, without needing to traverse the entire network segment to scan for devices. Existing port scanning methods generate a large number of probe packets, consuming internal network bandwidth and interfering with local large-scale models and AI agent inference operations. Furthermore, multiple gateways of the same model cannot be distinguished by port number alone, easily leading to accidental operation of others' private office AI workflows. This invention relies on a unique hardware SN to bind device identity, eliminating additional scanning traffic and enabling multi-role permission isolation to protect confidential local models and office data for government and enterprise use.

[0019] Furthermore, the offline caching fault tolerance feature ensures uninterrupted operation of private intranet AI services even when the external network is interrupted and the official website is inaccessible. The client reads the locally cached SN-IP mapping record, allowing it to still log in to the computing gateway via the SN to debug personalized workflows and run local large models within the local area network. Existing technologies rely entirely on real-time background queries and lack local caching, making device access completely impossible after an external network outage. This invention's offline adaptation feature perfectly meets the high-reliability private computing power service requirements of government and enterprise intranets.

[0020] Finally, the dynamic and automatic synchronization of mapping relationships adapts to DHCP dynamic IP environments, eliminating the need for manual maintenance of IP ledgers. The device periodically checks network and IP status every minute, and automatically updates the official mapping database after IP changes due to router restarts or device power outages and reconnections, without requiring manual table modifications or repeated network segment scans. Compared to manual static ledgers and static SN mapping solutions, this completely solves the pain points of delayed IP updates, login failures, and the need to check each device individually for lost ledgers, making it suitable for large-scale dynamic networking scenarios involving government and enterprise devices. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart of a login method based on an edge-side AI computing power device SN in a specific embodiment of the present invention; Figure 2 This is a schematic diagram of the login system based on the edge-side AI computing power device SN in a specific embodiment of the present invention; Figure 3 This is a schematic diagram of the overall process of the login method based on the edge AI computing power device SN in a specific embodiment of the present invention. Detailed Implementation

[0023] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0024] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0025] like Figure 1 As shown, Figure 1 This is a flowchart of a login method based on an edge-side AI computing power device SN in a specific embodiment of the present invention. The login method based on the edge-side AI computing power device SN includes: Step 1: The computing power gateway periodically executes network detection tasks and determines the network status through a two-layer connectivity check. Specifically, this includes: Step 11: The computing power gateway starts network testing 1 minute after powering on, and then performs a round of network verification every 1 minute thereafter. Step 12: The computing power gateway reads the device's unique hardware serial number and the local area network IP address of the local machine; Step 13: The computing power gateway executes a two-layer connectivity verification logic to avoid misjudgments of network status caused by ICMP (Internet Control Message Protocol) packets being blocked by the router. The two-layer connectivity verification includes the first layer of batch pinging public DNS nodes. If any node is successfully pinged, the network is considered normal. The second layer is a fallback interface verification. If all ping probes fail, the backend network probe interface is called to probe network connectivity (the purpose is to prevent networks blocked by the ping protocol). If both layers fail, the device is determined to have an abnormal network, the network is disconnected, and the hotspot startup process in Step 2 is initiated.

[0026] Step 2: When the network is determined to be disconnected, enable a temporary hotspot using an Access Point (AP) to achieve remote network configuration without a monitor. Specifically, this includes: Step 21: After the network detection determines that the connection is lost, the initialization process is executed to verify the resident virtual network interface card. This resident virtual network interface card is not deleted throughout the process to ensure a quick switchover when the network is restored. Step 22: Configure the parameters for the AP temporary hotspot on the computing power gateway; Step 23: Activate AP temporary hotspot connection. Users can directly scan and connect to the AP temporary hotspot using their mobile phones or laptops without needing any external devices to connect to the computing power gateway. Step 24: After the network configuration terminal connects to the AP temporary hotspot, it accesses the local network configuration page, fills in the network configuration parameters such as the target LAN WiFi name and password for the office / home, and sends them to the computing power gateway and saves them.

[0027] Step 3: Upon completion of network configuration or network recovery, the computing power gateway switches to STA mode and reports the SN and IP address differently. Specifically, this includes: Step 31: After the network configuration parameters are issued, the computing power gateway shuts down the AP temporary hotspot, enables STA mode, and loads the WiFi credentials within the computing power gateway to access the private local area network. Step 32: Perform the two-layer network connectivity verification again. If the external network is still unreachable, the computing power gateway will automatically switch back to the AP temporary hotspot and continue to retain the temporary network configuration channel. Step 33: After the network connection is stable, read the last reported IP (LAST_IP_CONF) cached in the computing power gateway and compare it with the current real-time IP of the device. If the IP has not changed, no reporting operation is performed to reduce invalid requests to the backend interface. If the IP has changed or there is no historical IP cache for the first time, the device's fixed SN and LAN IP are reported to the AIDA management backend. Step 34: After receiving the data, the AIDA management backend stores the dynamic mapping relationship between SN and IP using SN as the primary key, and writes the expiration time to indicate the online validity period of the device, thus constructing a dynamic SN-IP mapping database; Step 35: After successful reporting, the latest IP address is written to the local cache (LAST_IP_CONF) as the benchmark for the next round of comparison.

[0028] Step 4: The AIDA client queries the backend mapping data based on the SN to complete IP-free login. Specifically, this includes: Step 41: Office and home users with multiple roles open the local AI client and only need to enter the target computing power gateway hardware SN number. There is no need to manually search for or remember the IP address. Step 42: The AIDA client sends an address query request to the AIDA management backend, carrying the SN identifier; Step 43: The AIDA management backend retrieves the dynamic SN-IP mapping database, verifies the SN's validity and the device's online status, and returns the latest LAN IP bound to the SN after successful verification. Step 44: The AIDA client automatically fills in the resolved IP address, actively initiates a TCP remote connection, and directly logs into the computing power gateway; Step 45: After successful login, users can directly access the local large model, debug AI agents, and edit personalized workflows. Different user roles can isolate private AI business data by binding permissions according to SN.

[0029] In a specific embodiment of the application of the present invention, such as Figure 2As shown, the system of this invention comprises four main layers: an AI edge computing gateway 201, a network configuration terminal (mobile phone / laptop) 202, an AIDA management backend 203, and an AIDA client 204. The AI ​​edge computing gateway 201 has a built-in NPU for local large-model inference, and is equipped with a wireless network card, the auto_hotspot.sh script, a systemd service + timer scheduling unit, the NetworkManager network management tool, and a local configuration file that stores WiFi parameters and historical IPs. The AIDA management backend 203 includes two sets of backend API interfaces (network connectivity detection interface and device SN-IP reporting and storage interface). The AIDA client 204 can run on the network configuration terminal (mobile phone / laptop) 202.

[0030] Data flow sequence: AI edge computing gateway 201 performs periodic network detection → automatically starts AP hotspot when network is disconnected → network configuration terminal 202 connects to hotspot and sends WiFi configuration information → AI edge computing gateway 201 switches to STA mode to access the local area network → AI edge computing gateway 201 reads local SN and current IP → only when IP changes, reports SN-IP mapping relationship to AIDA management backend 203 → AIDA client 204 enters SN to query IP from AIDA management backend 203 → AIDA client 204 automatically fills in IP and establishes remote login channel.

[0031] The computing gateway locally stores two persistent configuration files: ~ / wifi_conf / wpa_sta.conf (stores office / home LAN WiFi account and password) and ~ / wifi_conf / lastip.conf (caches the last reported IP, used to determine whether to trigger the reporting logic); the gateway resides in the wlan_sta virtual wireless network card, which is not destroyed with each task, reducing network switching jitter.

[0032] In this embodiment, as Figure 3 As shown, the login method based on the SN of edge-side AI computing power devices specifically includes the following steps: Step 1: Systemd periodically triggers network detection tasks. This includes: Step 11: One minute after the computing power gateway is powered on, the auto-hotspot.timer timer starts the auto-hotspot.sh script for the first time, and then executes the full-process network detection logic once every 1 minute. Step 12: The script runs with root privileges and automatically reads the device's unique hardware serial number (obtained by calling the get_device_sn function, which serves as the device's unique identifier) ​​and calls get_device_ip to read the local area network IP address. Step 13: Execute a two-layer connectivity verification logic to avoid misjudgments of network status caused by ICMP packets being blocked by routing: The first layer involves batch pinging public DNS nodes: sequentially probing 223.5.5.5, 119.29.29.29, and 1.1.1.1, configuring the number of probes (PING_COUNT) and the timeout (PING_TIMEOUT). If any node is successfully pinged, the network is considered normal. The second layer is a fallback interface verification: if all ping probes fail, a GET request is initiated to call the backend network probe interface / api / v1 / device / network / check to probe network connectivity (this is to prevent pinging networks that are blocked). If the interface returns {"ok":true}, the network is considered normal. If both layers fail, the device's current network is considered abnormal, and the hotspot startup process begins. If the current network is normal, the process directly proceeds to step 33.

[0033] This invention relies on systemd managed script tasks. After the gateway restarts or the process crashes, the timer automatically resumes the periodic network detection logic, eliminating the need for manual program restart.

[0034] Step 2: In case of network anomalies, automatically enable AP hotspot for remote network configuration without a monitor. This includes: Step 21: After the network detection determines that the network is disconnected, the init_prepare initialization process is executed to verify the resident virtual network interface wlan_sta. If it does not exist, it is created automatically. The virtual network interface is not deleted throughout the process to ensure a fast switchover when the network is restored. Step 22: The script automatically creates an AP connection configuration for AIDA-AP through NetworkManager, with fixed parameters: working mode AP, local network segment 192.168.4.1 / 24, SSID (Service Set Identifier) ​​fixed as AIDA-AP, and open passwordless security policy; Step 23: Automatically activate hotspot connection. Users can directly scan and connect to the AIDA-AP temporary hotspot using their mobile phones or laptops without needing any external devices to connect to the computing power gateway. Step 24: After the network configuration terminal connects to the hotspot, it accesses the local network configuration page, enters the target LAN WiFi name and password for the office / home, sends it to the computing power gateway, and writes it to wpa_sta.conf for persistent storage.

[0035] Step 3: Network configuration complete / network restored, automatically switch to STA mode and report SN and IP addresses differently. This includes: Step 31: After the network configuration parameters are issued, the script shuts down the aida-ap hotspot, enables the STA to connect to aida-sta, and loads the WiFi credentials in wpa_sta.conf to access the private LAN. Step 32: Perform the two-layer network connectivity verification again. If the external network is still unreachable, automatically switch back to the AP hotspot and continue to retain the temporary network configuration channel. Step 33: After the network connection is stable, read the last reported IP cached in lastip.conf and compare it with the current real-time IP of the device: if the IP has not changed, no reporting operation is performed to reduce invalid requests to the backend interface; if the IP has changed, or if there is no historical IP cache for the first time, assemble a JSON reporting message: {"gateway":"device SN","host":"current LAN IP"}; Step 34: Call the backend IP reporting interface / api / v1 / device / ip / report via POST request. After receiving the request, the backend stores the dynamic mapping relationship between SN and IP with SN as the primary key, and writes the expiration time to indicate the online validity period of the device. Step 35: After successful reporting, the latest IP address is written to the local cache of lastip.conf and used as the benchmark for the next round of comparison.

[0036] Step 4: The AIDA client queries the backend mapping data based on the SN to automatically complete IP-free login. This includes: Step 41: Office and home users with multiple roles open the local AI client and only need to enter the target computing power gateway hardware SN number. There is no need to manually search for or remember the IP address. Step 42: The AIDA client sends an address query request to the AIDA management backend, carrying the SN identifier; Step 43: The AIDA management backend retrieves the dynamic mapping database, verifies the SN's validity and the device's online status, and returns the latest LAN IP bound to the SN after successful verification. Step 44: The AIDA client automatically fills in the resolved IP address, actively initiates a TCP remote connection, and directly logs into the computing power gateway; Step 45: After successful login, users can access various private AI services on the AI ​​computing power gateway, such as local large models, debugging AI agents, and editing personalized workflows, based on their different user roles and permissions.

[0037] Therefore, the operation and maintenance client can initiate an IP query to the server by simply entering the device SN. The server performs multi-level verification of SN validity, whitelist, and device online and returns the IP. The client automatically fills in the IP and establishes a TCP remote login channel. Based on the hardware SN, multi-role access permission isolation is achieved. Different users can only access the computing power gateway corresponding to their own bound SN, thus isolating their personalized AI workflows and local large model private data.

[0038] In one embodiment, the AI ​​client locally caches recent SN-IP mapping records. When the external network is disconnected or the backend service is inaccessible, it directly reads the local cache to complete SN addressing login within the local area network, thus achieving offline cache fault tolerance. The backend periodically scans the mapping library and clears offline device records that have exceeded their validity period to avoid returning invalid IPs and causing login failures, thus implementing an offline device cleanup mechanism.

[0039] This invention targets distributed AI edge computing gateways within private LANs in offices and homes. The device locally carries a large model and a built-in AI agent, supporting multi-role user customization of personalized AI workflows. The entire solution relies on the NetworkManager network management tool of the Linux system, the auto_hotspot.sh core script, the systemd timed scheduling service, and the backend API reporting interface to build a complete automated system. No external monitor or keyboard and mouse are required for local operation of the device. It achieves automatic hotspot network configuration, intelligent network status switching, automatic reporting of SN binding to IP, and client login to the computing gateway without IP by entering the SN. It includes a complete execution process, hardware and software collaborative architecture, and core execution logic.

[0040] This invention adopts a differentiated SN-IP dynamic reporting mechanism. The gateway caches historical IPs locally and only reports the SN and LAN IP to the server when it goes online for the first time or when the IP changes. It performs periodic checks every 1 minute, which greatly reduces the frequency of interface requests and server load. It automatically maintains the real-time updated device address mapping relationship and is compatible with DHCP dynamic IP networking environments.

[0041] This invention adopts an automatic device discovery and IP-free login architecture based on a unique hardware serial number (SN). It abandons LAN port scanning and manual ledger registration methods, and uses the factory-fixed SN of the device as a unique index. The client only needs to enter the SN to resolve the device's internal network IP and automatically establish a login session. Relying on the SN, it realizes multi-role hierarchical authentication and data isolation, avoiding confusion between multiple computing power gateways of the same model and leakage of private AI business data.

[0042] This invention adopts an external network offline caching fault-tolerant addressing scheme. The client caches the SN-IP mapping record locally. When the external network is interrupted and the cloud management service cannot be accessed, the local area network can still log in to the computing power gateway by resolving the IP through the SN, ensuring the uninterrupted operation of local large model and AI intelligent agent private computing power services.

[0043] This invention adopts a standardized hardware and software architecture, with the entire logic solidified and integrated into the AIDA computing power gateway. It is equipped with a cloud management interface from the official website and can be commercially delivered in conjunction with terminal hardware and software. The entire set of scripts, systemd timed scheduling, and interface message formats are standardized, enabling large-scale batch deployment in government and enterprise private office scenarios.

[0044] In a specific embodiment of this invention, a local edge server can replace the cloud-based official website (imaida.cn) for SN-IP mapping storage, eliminating the unified SN-IP mapping library on the public cloud website. A separate local edge management service program is deployed within the government / enterprise / home LAN. After the computing power gateway completes network configuration, it no longer sends data to the public network POST reporting interface, but instead reports the SN and LAN IP to the internal network edge service. In this solution, the computing power gateway's auto_hotspot.sh script only modifies the internal network address of the reporting interface; the entire logic of dual-layer network detection, automatic AP / STA switching, and hotspot network configuration remains completely unchanged. The AIDA client and supporting terminals no longer access public domain names to query IPs, but only initiate SN addressing requests to the internal network edge service. In offline scenarios, the internal network edge service continues to run, and SN resolution and login can be completed without relying on the client's local cache. This solution is suitable for private deployment scenarios in confidential government and enterprise environments where the internal network is completely isolated and devices are prohibited from accessing the external network, achieving all the invention's effects of automatic network configuration, SN addressing, and automatic device discovery.

[0045] In a specific embodiment of this invention, Bluetooth or QR code local reading of the SN replaces manual SN input. This solution retains the core processes of automatic hotspot network configuration, periodic network detection, SN-IP cloud reporting, and backend mapping storage, only optimizing the client-side SN input interaction, eliminating the purely manual SN input method. In this solution, a device SN QR code is printed on the AIDA computing power gateway. Tablets and maintenance clients are equipped with QR code scanning modules to automatically read the SN by scanning the QR code; alternatively, the gateway has a built-in Bluetooth broadcast module, allowing clients to scan the nearest device SN via Bluetooth and automatically fill it into the query box. The underlying core technologies such as network switching, IP reporting, SN authentication, and offline caching remain unchanged; only the front-end interaction method is replaced. This solution reduces numerical errors caused by manual SN input and improves operational convenience, while the underlying core addressing and network configuration mechanisms still fall within the scope of this patent protection.

[0046] In a specific embodiment of this invention, a wired Ethernet local network configuration replaces the Wi-Fi hotspot AP configuration. When a device loses network access, the wireless Wi-Fi hotspot is not activated; instead, a temporary wired network port is reserved as the local network configuration channel. In this scheme, the network detection logic, SN+IP reporting mechanism, and SN resolution login process are entirely based on this invention. After network failure is detected, the temporary wired network port is automatically activated, allowing users to connect devices such as laptops using network cables to complete the local area network (LAN) WiFi parameter distribution and configuration. After configuration is complete, the connection automatically switches back to the main wired / wireless STA link, replacing only the wireless AP configuration with the temporary wired network configuration channel. This scheme is suitable for confidential computer rooms without wireless environments or with wireless signal shielding, still achieving the invention's objective of remote network configuration without a monitor peripheral and automatic device discovery and login.

[0047] In summary, this invention employs a dual-layer network connectivity detection + AP / ST intelligent switching integrated mechanism, using public network DNS multi-node ping detection + cloud interface verification dual-layer judgment logic to solve network status misjudgments caused by ICMP blocking; it automatically opens a passwordless temporary hotspot when the network is disconnected and automatically closes the hotspot and switches back to STA mode when the network is connected. With a resident virtual network card, it reduces network switching jitter. The entire process can be completed remotely without a monitor, keyboard and mouse, making it suitable for batch deployment scenarios without peripherals.

[0048] Accordingly, this invention also provides a login system based on the SN of an edge-side AI computing power device. This login system includes the following modules: an AI edge computing power gateway with a wireless network card, a network configuration terminal (mobile phone or laptop), an AIDA management backend, and an AIDA client. The AI ​​edge computing power gateway is equipped with a network management script that is executed periodically. It achieves automatic Wi-Fi hotspot activation when the network is disconnected through dual-layer connectivity detection to complete remote network configuration without peripherals. After the network configuration is connected to the intranet, it only actively reports the device's fixed SN and local area network IP to the AIDA management backend when the IP changes. The AIDA management backend builds a dynamic SN-IP mapping library with the SN as the primary key. The AIDA client can query the corresponding device IP from the AIDA management backend by entering the SN and automatically establish a remote login channel. At the same time, it is equipped with a local caching mechanism for the AIDA client. When the external network is disconnected, the intranet can still rely on the cache to achieve SN addressing and login. It is suitable for office, home local large model, and AI intelligent agent multi-role private computing power service scenarios.

[0049] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0050] The present invention also provides a storage medium, which is a computer-readable storage medium storing a computer program thereon, the computer program being executable when it is run. Figure 1 The method shown may include some or all of the steps. The storage medium may include read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc. The storage medium may also include non-volatile memory or non-transitory memory, etc.

[0051] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data provider to another website, computer, server, or data provider via wired or wireless means.

[0052] The embodiments of the present invention have been described in detail above. Specific implementation methods have been used to illustrate the present invention. The descriptions of the embodiments above are only for the purpose of helping to understand the methods and systems of the present invention, and are merely some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention, and the content of this specification should not be construed as a limitation of the present invention. Therefore, any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A login method based on an edge-side AI computing power device SN, characterized in that, The method includes: Step 1: The AI ​​edge computing gateway periodically performs network detection tasks and determines the network status through two-layer connectivity verification. Step 2: When the network is determined to be disconnected, the AI ​​edge computing gateway opens a temporary AP hotspot to perform remote network configuration of the network configuration terminal without a display. Step 3: When the network configuration is completed or the network is restored, the AI ​​edge computing power gateway switches to STA mode and reports the device's fixed SN and LAN IP to the AIDA management backend in a differentiated manner. Step 4: The AIDA client queries the mapping data between the device's fixed SN and the local area network IP in the AIDA management backend based on the device's fixed SN of the AI ​​edge computing gateway, and completes the IP-free login to the AI ​​edge computing gateway.

2. The login method based on the edge-side AI computing power device SN according to claim 1, characterized in that, Step 1 includes: Step 11: The AI ​​edge computing gateway starts network detection 1 minute after powering on, and then performs a round of network verification every 1 minute thereafter. Step 12: The AI ​​edge computing gateway reads the device's fixed SN and reads the local area network IP; Step 13: The AI ​​edge computing gateway performs a two-layer connectivity check, which includes: a first layer of batch pinging of public DNS nodes; if any node is successfully pinged, the network is considered normal; a second layer of fallback interface check; if all ping probes fail, an attempt is made to access the network probe interface of the AIDA management backend to probe network connectivity; if both layers fail, the current network is considered abnormal, the network is disconnected, and the process proceeds to step 2.

3. The login method based on the edge-side AI computing power device SN according to claim 1, characterized in that, Step 2 includes: Step 21: The AI ​​edge computing gateway sets the parameter information for the AP temporary hotspot; Step 22: The AI ​​edge computing gateway activates the AP temporary hotspot connection, and the user uses the network configuration terminal to scan and connect to the AP temporary hotspot; Step 23: After the network configuration terminal connects to the AP temporary hotspot, it accesses the network configuration page of the AI ​​edge computing power gateway, fills in the network configuration parameters such as the target local area network WiFi name and password for the office or home, and the AI ​​edge computing power gateway saves and starts the newly configured network.

4. The login method based on the edge-side AI computing power device SN according to claim 3, characterized in that, Step 3 includes: Step 31: After the network configuration parameters are issued, the AI ​​edge computing power gateway shuts down the AP temporary hotspot, enables STA mode, and loads the WiFi credentials within the AI ​​edge computing power gateway to access the private local area network. Step 32: The AI ​​edge computing gateway performs a two-layer connectivity check again. If the external network is still unreachable, the AI ​​edge computing gateway automatically switches back to the AP temporary hotspot and continues to retain the temporary network distribution channel. Step 33: After the network is stably connected, the AI ​​edge computing power gateway reads the previously reported local area network IP cached in the AI ​​edge computing power gateway, compares it with the real-time local area network IP of the current distribution terminal, and reports the device fixed SN and local area network IP of the AI ​​edge computing power gateway to the AIDA management backend. Step 34: After receiving the data, the AIDA management backend stores the dynamic mapping relationship between the device's fixed SN and the local area network IP address of the AI ​​edge computing gateway using the device's fixed SN as the primary key. At the same time, it writes the expiration time to indicate the device's online validity period and constructs a dynamic SN-IP mapping database. Step 35: After the AI ​​edge computing gateway successfully reports the latest local area network IP address, write the latest IP address into the AI ​​edge computing gateway cache as the benchmark for the next round of comparison.

5. The login method based on the edge-side AI computing power device SN according to claim 4, characterized in that, In step 33, if the local area network IP remains unchanged, the AI ​​edge computing gateway does not perform a reporting operation; if the local area network IP changes, or if the cache of historical local area network IPs is not started for the first time, the AI ​​edge computing gateway reports the device's fixed SN and local area network IP to the AIDA management backend.

6. The login method based on the edge-side AI computing power device SN according to claim 4, characterized in that, Step 3 also includes the AIDA management backend periodically scanning the dynamic SN-IP mapping database and clearing offline device records that have exceeded their validity period.

7. The login method based on the edge-side AI computing power device SN according to claim 4, characterized in that, Step 4 includes: Step 41: The user opens the AIDA client and enters the device's fixed SN number as the target AI edge computing gateway; Step 42: The AIDA client first queries the historical login address of the AI ​​edge computing power gateway recorded locally; if the AIDA client cannot log in to the historical login address, it sends an address query request to the AIDA management backend carrying the device's fixed SN of the AI ​​edge computing power gateway. Step 43: The AIDA management backend retrieves the dynamic SN-IP mapping database, verifies the SN's legality and the device's online status, and returns the latest LAN IP bound to the SN after successful verification. Step 44: The AIDA client uses the latest LAN IP address obtained from the query to directly log in to the AI ​​edge computing gateway that is the login target; Step 45: After successful login, users can access the local large model, debug AI agents, and edit personalized workflows of the AI ​​computing power gateway according to their different role user permissions.

8. A login system based on an edge-side AI computing power device (SN), characterized in that, The system includes an AI edge computing gateway, an AIDA client, an AIDA management backend, and a network configuration terminal. The AI ​​edge computing gateway periodically performs network detection tasks and determines the network status through a two-layer connectivity check. When the network is determined to be disconnected, the AI ​​edge computing gateway activates a temporary AP hotspot to perform remote network configuration for the network configuration terminal without a monitor. When network configuration is completed or the network is restored, the AI ​​edge computing gateway switches to STA mode and reports the device's fixed SN and LAN IP address to the AIDA management backend in a differentiated manner. The AIDA client queries the mapping data between the device's fixed SN and LAN IP address in the AIDA management backend based on the device's fixed SN to complete IP-free login.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program, when run by the processor, executes the steps of the login method based on the edge-side AI computing power device SN as described in any one of claims 1 to 7.

Images