A Security-enhanced Advertising Platform based on Blockchain and Edge Computing in Generative AI

ABSTRACT

With the development of advertising technology, especially the emergence of generative artificial intelligence, the generation and editing of advertising content has become more convenient and efficient. However, this also brings new security risks to digital advertising. For example, in order to generate personalized advertising content, AI can collect user behavior data without the user’s knowledge and automatically generate ads that are beneficial to advertisers. In addition, the main issue regarding outdoor advertising is whether the outdoor advertising platform can deliver the advertising content required by the advertiser at the designated place and time according to the actual requirements of the customer. Combining blockchain and edge computing technology, this paper builds a security-enhanced outdoor advertising platform. The decentralized, non-tamperable and traceable features of blockchain technology provide the advertising platform with credibility between users and advertisers, as well as credibility between advertisers and the advertising platform. At the same time, the terminal controller with edge processing capabilities improves the efficiency of advertising and enhances the task balance between the server and the playback side. Experimental results show that the platform can effectively manage user identities and maintain advertising plans. It not only improves data security and credibility, but also has good performance.

Introduction

Digital advertising is one of the key drivers of the digital economy, and out-of-home digital advertising is characterized by mandatory viewing, precision and focused delivery, which makes advertising investment more valuable. It is estimated that the cost of advertising investment will reach a staggering $646 billion by 2024 (Statista 2023). Along with the increase in advertising costs, digital advertising technology and delivery formats have evolved from the initial banner advertising to the current mobile advertising and computational advertising. With the development of advertising technology as well as the advertising industry, especially the emergence of generative artificial intelligence, it has become easier and more efficient to generate, edit and deliver digital advertising content. However, this has also brought new security risks to the advertising industry (Yun and Strycharz 2023).

With the birth and widespread application of generative AI technology, the emergence of intelligent synthetic media has further increased the risk of content being modified and edited, thereby further increasing users’ distrust of advertising information (Baek 2023; Pataranutaporn et al. 2021). Specifically, the necessary trust between users and advertisers is not established. At the same time, in the ecological chain of advertising, it is difficult to trace who made the ads, who placed a particular ad, and when and where it was placed, and for which groups of people. In addition to the challenges of information security, energy consumption is also an important issue in the advertising delivery process. Currently, there are various advertising delivery methods on the market, and the display terminals represented by LED screens require a large amount of electrical support, which has an impact on the environment. Moreover, LED screens take up a lot of space and have high maintenance costs.

In response to the issues related to digital advertising discussed above, we carefully analyze the disadvantages of the current advertising model, especially the security threats of outdoor advertising that rely on centralized management and delivery, and try to adopt a new way to deal with these problems. In this paper, a decentralized security-enhanced outdoor advertising platform is constructed by introducing blockchain technology and edge computing technology. Based on the decentralized, untamperable, and traceable features of blockchain technology, the advertising platform provides trustworthiness between users and advertisers. Meanwhile, through the smart contract deployed on the blockchain, the authentication of user identity, the dynamic generation of advertisement delivery plan and the verification of advertisement delivery content are automatically accomplished, which establishes the trustworthiness between advertisers and the advertising platform. In addition, on the delivery side of the platform, edge nodes are formed through edge controllers, and dimming glass is introduced as a playback medium, which not only reduces the computational pressure on the blockchain network and improves the load balancing of the platform, but also increases the flexibility of outdoor advertising delivery.

This paper’s contributions are as follows:

• We propose a decentralized and security-enhanced lightweight advertising platform. Users can remotely access the platform through client nodes to achieve remote delivery of advertisements.

• By using blockchain technology, the platform realizes the identity authentication of the system and the decentralized management of data, which ensures the data security of advertisers during the process of advertising placement. Combined with the programmable characteristics of smart contracts, it completes the writing of the main business of the system and simplifies the development of the application layer.

• By leveraging edge computing technology, we have independently developed edge control nodes, thereby alleviating the computational burden on the blockchain network. The playback terminal uses a combination of dimming glass and projectors to deliver ads in a novel and flexible way. Experiments show that the system has excellent advertising performance.

Background and Related Technologies

Blockchain

Blockchain is a distributed database technology used to securely record transaction information without the need for third-party intermediaries such as banks and governments. A blockchain consists of blocks, with the first block known as the genesis block. Transaction data is stored in blocks in encoded form, and each block uses hash value as a unique identifier (Peres et al. 2023). When creating a new block, a hash operation is performed on all the information in the current block to obtain a unique hash value. Once a block is created, any changes to the contents of the block will result in a change in the hash value. By linking blocks through hash values, an immutable and continuously growing chain is formed (Cao et al. 2024; Miao et al. 2024). In recent years, blockchain has gained significant attention in various application areas such as supply chain management, digital identity verification, and IoT security. A study suggests that blockchain will bring various opportunities to virtual worlds like the metaverse, triggering a new wave of technological innovation and industrial transformation (Yang et al. 2022).

Hyperledger Fabric, as a permissioned blockchain, has been widely used in enterprises. Fabric provides highly modular services such as chaincode, membership service providers (MSP), consensus algorithms, and state databases. Chaincode as smart contract defines transaction rules, MSP handles all encryption technologies such as identity management, signature generation, and verification. Consensus algorithms are used to validate and update transactions in the ledger, and state databases include Level DB and Couch DB (Sharma, Jindal, and Dutta Borah 2024).

A study (Qiao et al. 2023) presented a decentralized privacy-preserving credit assessment system called PEevaChain based on Hyperledger Fabric blockchain, which ensures trusted data and computation while leveraging Hyperledger Fabric’s access control mechanisms to prevent unauthorized access. Another study (Pelekoudas-Oikonomou et al. 2023) integrated blockchain technology into a new solution for IoMT networks, using Hyperledger Fabric blockchain to provide distributed database storage, eliminate single points of failure, enhance the security of IoMT medical monitoring systems, and deploy a blockchain security architecture. Additionally, researchers (Pulmano et al. 2023) developed a digital credential platform based on Hyperledger Fabric, which has been validated in national identity recognition systems and academic credential systems. They developed chaincodes for each type of digital credential, and deployed them on specified channels in the network. Literature (Liang et al. 2023) builds a highly available blockchain platform for education alliance, which uses kubernetes to complete Hyperledger Fabric network deployment and uses HotStuff consensus algorithm to access sorting services. At the same time, in order to improve the resource utilization of the chaincode, the author manages the chain code through the functional computing service. In addressing the challenges of managing access permissions for IoT devices, a study (Zaidi et al. 2023) proposed an IoT security framework using Hyperledger Fabric and role-based access control policies. The framework incorporates a role management system that resolves conflicts based on predefined rules and user preferences, combined with a consensus mechanism to determine user’s access rights.

In the security-enhanced advertising platform, we use Hyperledger Fabric to store ad information and use its transparent, decentralized characteristics to ensure data security. At the same time, combined with the identity management mechanism, we complete the edge node authentication so as to ensure the security of communication.

IPFS

IPFS is a distributed file system, with a non-centralized peer-to-peer file transfer mode, and uses a content-addressed file transfer protocol (Dwivedi, Amin, and Vollala 2023). When uploading a file, to ensure the file is not tampered with, IPFS calculates hash value of this file and uses this value as the unique identifier for the file. The main technologies of IPFS include Distributed Hash Table (DHT), Merkle Directed Acyclic Graph (DAG), and version control (Li et al. 2023). DHT (Trautwein et al. 2022) provides IPFS with distributed storage and retrieval capabilities, enabling fast location and retrieval of files within distributed nodes. Merkle DAG (Jayabalan and Jeyanthi 2022) is used to verify the integrity and correctness of nodes, supporting content addressing in IPFS.

During the process of remote advertising delivery, the advertisement files may range from tens to hundreds of megabytes. Although blockchain can be used for file storage and management, it also has certain limitations.

In terms of storage performance, to ensure data immutability, blockchain data is redundantly backed up to each node, and the storage capacity of the blockchain is limited by each node, resulting in high storage costs. IPFS is a decentralized distributed file system, and its storage capacity depends on the total storage capacity of all nodes in the network. IPFS divides files into small chunks and stores them in a decentralized manner across the network, with nodes only needing to store file fragments without backing up the entire file, reducing storage costs (Benet 2014). In terms of data transmission, data transmission in blockchain requires consensus and verification from multiple nodes, resulting in relatively slow transmission speeds. IPFS utilizes content addressing and caching mechanisms to quickly retrieve files that already exist on the network. Additionally, due to IPFS’s peer-to-peer network structure, files can be directly retrieved from nearby nodes, reducing transmission latency. Therefore, we have adopted an off-chain storage solution.

We store the advertisement files in a decentralized manner on nodes in the IPFS network. After the terminal app retrieves an advertisement file, it calculates hash A based on the file information and then compares hash A with hash B stored on the chain to verify whether the obtained file matches the target file.

MQTT

MQTT is a lightweight communication protocol based on the TCP/IP protocol, released by IBM in 1999. MQTT adopts publish/subscribe mode. Message senders publish messages to topics, and devices or applications subscribing to these topics will receive these messages. This protocol is mainly used for machine-to-machine (M2M) communication (Kegenbekov and Saparova 2022). The content of MQTT messages mainly consists of a fixed header, a variable header, and a payload. The maximum transmission size can be extended to 256MB (Alshammari 2023).

A study (Boppana and Bagade 2022) found a specific vulnerability in unencrypted IoT systems based on MQTT, allowing attackers to launch man-in-the-middle attacks and XSS attacks to unauthorized access the entire system. To address the weak security of MQTT, a study (Akshatha and Dilip Kumar 2023) proposed a solution to enhance MQTT communication security using blockchain sharding, in which each shard operates as an independent chain. By combining with a consensus mechanism, this solution can improve security while minimizing computational overhead. Another study (Zeghida, Boulaiche, and Chikh 2023) proposed an intrusion detection model based on ensemble learning, where authors independently trained several models on random subsets of the dataset and averaged the predictions to calculate the final result, thereby improving the accuracy of intrusion detection results.

In our advertising platforms, the communication overhead between the terminals and edge control nodes needs to be ensured is as low as possible. By integrating the MQTT protocol into the platform, we reduce the communication overhead between the edge control nodes and the terminals while ensuring real-time access to device status.

Edge Computing

In traditional advertising delivery, the ad files need to be transmitted from the server to the terminal devices, which may result in untimely advertisement broadcasting if only relying on server-side processing. Edge computing, on the other hand, is a distributed computing model that arranges computing resources around the device. Its goal is to perform data processing in a location closer to the user or device to reduce latency and improve efficiency (Hua et al. 2023; Walczak et al. 2023).

To address the issue of highly dispersed edge computing devices and the inability to identify malicious nodes, literature (Liao and Cheng 2023) suggested using blockchain to enhance network security and proposed a reputation and vote-based consensus mechanism that does not require complex hash calculations. At the same time, to prevent malicious nodes from participating in the consensus, the authors designed a filtering algorithm for detecting and filtering malicious nodes. In order to improve environmental monitoring applications, literature (Roostaei et al. 2023) proposed integrating Internet of Things sensors and edge computing and designed two pilot applications to compare data latency and energy consumption the IoTEC approach and the conventional sensor monitoring method. Literature (Rivera, Refaey, and Hossain 2020) used blockchain to establish a trusted collaborative mechanism among edge nodes in a multi-access edge computing environment, enabling the sharing of tasks among nodes and achieving resource sharing of edge computing devices.

In our platform, we develop an edge control node located at the edge of the terminal devices, which is used to interact with the blockchain network to complete the binding with the terminal device and the issuance of plans. The edge control node adopts ARM chip and integrates PetaLinux operating system. In addition, the edge node also integrates the MQTT protocol to receive real-time device status.

Dimming Glass

To build green and smart homes and effectively control indoor lighting and temperature, researchers have invented dimming glass to efficiently control the sunlight entering buildings (Chang et al. 2022). Dimming glass is a new type of glass product that controls transparency through voltage. It features switchability, energy efficiency, privacy protection, and is widely used in architectural glass (Nundy et al. 2021). Traditional LED screens used in advertising have high brightness and energy consumption, typically for outdoor use, which can lead to eye discomfort after long-term viewing. Compared to LED screens, dimming glass can flexibly adjust brightness, significantly reduce energy consumption, and is suitable for places like offices and meeting rooms that require flexible light control and privacy (Iluyemi et al. 2022).

Threat Model

Traditional digital advertising delivery adopts the structure of management-server-terminal. Management is used by advertisers to access the application services provided by server, such as data storage service and web application service. Through web application, advertisers can set up playback files for terminal devices. Server communicates with management and terminal. It is also responsible for storing ad information uploaded by users and sending playback plans to terminal devices. Terminal refers to the final display device for advertising delivery. Usually, the terminal device installs an app for communicating with the server and obtaining the ads for playback.

As shown in Figure 1, the traditional advertising model has the following problems:

Figure 1. Traditional advertising model.

1. In the advertising platform, the server is the core of the system, which adopts centralized data storage services such as databases, and is easy to suffer from SQL injection, DDOS, and single point of failure attacks. Once it collapses, it will lead to platform paralysis and data security is not guaranteed.

2. With the increase of data and connections, the server faces huge computational and storage pressure, resulting in slow response, high latency, and poor user experience.

3. Attackers take advantage of app loopholes in terminal devices to tamper with the advertisement information received by the terminal, which can result in inconsistency between the advertiser’s planned content and the broadcast content.

In response to threat 1, we adopt decentralized distributed storage to store data files to reduce the risk of single-point attack and ensure data security. In addition, it also enhances system security and robustness.

For threat 2, we independently developed an edge control node, which is used to reduce the computational pressure on the server side and solve the problems of slow response and low latency. The edge control node listens to changes in the account book to complete the plan issuance, monitor of equipment status in real time and control of terminal switches, which greatly shortens the data transmission distance. Additionally, the system utilizes IPFS for file storage, where files are stored in a decentralized manner across network nodes. Considering the issue of processing latency for large files, we adopt a nearby storage approach and introduce a caching mechanism. The system stores files on nodes close to the terminal device to reduce the data transmission distance. Moreover, the terminal devices integrate a caching module to cache the required files before playing advertisements, reducing file processing latency.

In response to threat 3, we design a resource verification algorithm in the terminal app. The app first obtains the hash value of the advertisement file from the blockchain network through the edge node, and then calculates the hash value of the received file using the MD5 algorithm. By verifying whether the two hash values are consistent, it determines whether the terminal advertisement has been tampered with.

As a result, we design a security-enhanced advertising platform that incorporates blockchain and edge computing.

System Design

Conceptual Model

The security-enhanced advertising platform consists of a network layer, an edge layer and a terminal layer. The network layer is divided into blockchain network and off-chain network. The blockchain network adopts Hyperledger Fabric, while the off-chain network uses IPFS. Peer nodes of the blockchain network are composed of cloud servers and smart gateways, which provide services such as data storage and identity authentication. Users use client nodes to access the network, and edge nodes are responsible for sending information to terminal devices. The off-chain network is responsible for storing large advertisement files. The terminal layer consists of a smart projector and a dimming glass. The smart projector has a built-in app for accessing the playback schedule stored in the network layer, and the dimming glass acts as a curtain and as a playback medium. The proposed conceptual model is shown in Figure 2.

Figure 2. Conceptual model.

Client nodes refer to the management that integrates with Hyperledger Fabric sdk. Client nodes allow users to set up playback schedules and manage devices. At the same time, users can verify whether the schedule information, such as playback files, playback time, etc., is consistent with the schedule they assigned, and confirm whether the information has been tampered with.

Network layer provides services such as data storage, smart contracts, and identity authentication. The identity certificates required for identity authentication are issued by certificate authority (CA) nodes, and the authentication algorithm is shown in Algorithm 2. Server in the network provides MQTT services. The blockchain network is capable of managing advertisement information, storing playback schedules, and monitoring device status in real time. Different advertisers use channels to segregate data. With the help of channels, advertisers can independently manage ad files and devices. Usually, a channel contains an advertiser and a service provider. By leveraging the decentralized nature of blockchain, data security can be enhanced, reducing the risk of data tampering.

Edge control nodes are used as edge client nodes to communicate with the blockchain. Edge control nodes are responsible for networking and device binding, and the binding algorithm is shown in Algorithm 3. It also controls the switch of terminal devices with Algorithm 1. Additionally, the edge nodes are responsible for uploading real-time status information of devices and sending playback schedules to terminal devices. It decreases the distance of data transmission and reduces the computational pressure of the blockchain network.

Terminal devices include a smart projector and a dimming glass. The smart projector and diming glass receive electrical signal control from the edge control node. The projector retrieves the playback schedule from the edge node. The dimming glass acts as the medium for advertising playback. Before playing, the projector will use Algorithm 4 to complete the resource verification.

Business Process

In the advertising platform, advertisers need to access the blockchain with certificate, and the certificate is issued by CA nodes. Users can complete the management of terminal devices through the blockchain network. An edge node controls a group of terminals, which are composed of a dimming glass and a smart projector. Business process is shown in Figure 3.

Figure 3. Business process.

Step 1: The authenticated users log in to the advertising platform using Algorithm 2. Since channels are used to isolate data from different advertisers, they can only maintain their own devices and ad files.

Step 2–4: Users upload advertisement files, set up playback schedules for playback terminals, and control the switch of the devices by calling smart contracts through the cli-based interface. The uploaded data, such as ad files, need to be audited by channel members, and unqualified advertisement files should be deleted. The information of approved files will be recorded on the blockchain, while the documents themselves will be stored in a decentralized off-chain network.

Step 5: The cli-based interface retrieves the information uploaded by the user, and the device status from the ledger by calling the smart contract, and then passes it to the user interface in the specified data format to render the data in real time.

Step 6–7: The edge control node listens for data changes in the ledger, such as device switches and schedule information, and queries the latest information. Since the query operation does not affect the state of the ledger, it does not need to be audited. The edge control node also integrates MQTT services and stores relevant information.

Step 8: The edge control node controls the terminal according to the device state. If it is the off state, the edge node will use Algorithm 1 to power up the terminal and send its cached configuration information, such as the MQTT communication address, to the smart projector.

Step 9–10: The edge control node receives the binding request from the projector, then it uses Algorithm 3 to bind its own device number with the projector’s number, and registers them on the blockchain network.

Step 11: The projector obtains the playback schedule from the edge control node. Based on the video download address in the schedule information, the projector downloads and caches the advertisement files from the off-chain network. At the same time, the projector uses Algorithm 4 to verify if the resource file matches the target file. To prevent unsuccessful requests, the projector sends multiple requests until the request is successful.

Step 12: Projector checks the obtained playback schedules, and then plays the advertisements within the specified schedule time. When all the advertisements within that schedule time have been played, it continues to play in a loop until a new schedule is notified.

System Implementation

Edge Control Node

As shown in Figure 4, the edge node can connect to the local area network via Ethernet or WIFI. In order to maintain the speed and stability of communication, we use the wired way to connect to the network. In addition, it controls the switching synchronization of the projector and dimming glass by receiving infrared signals. To support the permanent storage of data such as MQTT addresses, we add a Flash module. The DOUT module uses a relay, and the edge node supplies power to the projector and dimming glass through the relay.

Figure 4. Edge control node.

App in Projector

The projector app is used to receive the playback schedules stored on the blockchain. Once the projector is turned on, the app begins its operation. It first searches for nearby edge control nodes. After successfully searching, it obtains the configuration such as MQTT address, and then it requests for the playback schedules. Upon successfully acquiring the schedules, it begins playing them in a loop until a new plan is issued. The app workflow is shown in Figure 5.

Figure 5. App workflow.

System Network

A security-enhanced advertising platform consists of a dimming glass, a smart projector, an edge control node, a blockchain network, and an off-chain network. When the platform starts to work, the smart projector searches for the edge control node via the broadcast address, and then requests the configuration information from the edge control node after a successful search. At the same time, the edge control node will complete the binding with the device, and then register the binding information on the blockchain. The edge control node acquires the device status in real time. When the device status makes changes, it will report the change information. The projector app sends a request to the edge control node to obtain the schedule, and the edge node carries the certificate to obtain the schedules from the blockchain network, and then sends them to the smart projector. The projector caches the video file according to the video address in the schedules and verifies the hash value between the cached file and the target file, so as to determine whether the cached file is consistent with the target file. The system timing is shown in Figure 6.

Figure 6. System timing.

During the searching time, the projector uses the UDP to obtain the communication address of the edge control node. In the time of obtaining configuration, HTTPS is used to send a request to communication address of the edge node to get the MQTT address. In the device binding phase, the edge node uses gRPCS to communicate with the blockchain and register the binding information on the blockchain. The terminal device uses MQTT to report the status, and the edge node forwards the status with gRPCS. In the stage of requesting plan, the projector app uses HTTPS to send a request to the edge node, and the edge node uses gRPCS to forward the request, and after the request is successful, the download url of the advertisement file and its start and end time will be included in the response information. The specific data format is shown in Table 1.

Table 1. Message format.

Event Type	Data Format	Protocal
DiscoverRequest	{“serial_number:””,””ip:”””}	UDP
DiscoverResponse	{“serial_number:””,””url:”””}	UDP
SysSettingRequest	{“serial_number:”””}	HTTPS
SysSettingResponse	{“serial_number:””,””mqtt:””,””ip:”””}	HTTPS
BindRequest	{“projectorID:””,””controlboxID:”””}	gRPCS
DeviceStatus	{“serial_number:””,””curtainStatus:””on/off,” ”projectionStatus:””on/off”}	MQTT
ProjectorStatus	{“serial_number:””,””projectionStatus:””on/off”}	MQTT
PlanRequest	{“serial_number:”””}	HTTPS
PlanRequest	{“serial_number:”””}	gRPCS
PlanResponse	{“plans:”{“startTime:””,””endTime:””,””video:””url”}, ”serial_number:”””}	gRPCS
PlanResponse	{“plans:”{“startTime:””,””endTime:””,””video:””url”}, ”serial_number:”””}	HTTPS

The projector and dimming glass are bound one-to-one, controlled by infrared signals or remote control signals to operate their switches, and must maintain power synchronization.

Algorithm Design

Power Control

The edge control node uses a relay to power up the devices. The projector and the dimming glass are connected in parallel. When the control signal is received, it will turn on the relay. The current will flow through the output ports of the relay and power both devices simultaneously. Algorithm 1 is the power up pseudo-code.

Table

Algorithm 1: Power Control
INPUT: controlSignal OUTPUT: response if controlSignal = = 1 then if powerStatus= = 0 then power up a pair of devices powerStatus = 1 return response result else power down a pair of devices powerStatus = 0 return reponse result end if end if

Identity Authentication

The authentication pseudo-code is used to verify the user’s identity information. When accessing the platform, the user needs to provide the private key and certificate file. When the system receives the private key and certificate file, it will first judge the file type. Then fabric sdk will verify the legitimacy of the certificate. After verifying successfully, it will save the information to the wallet folder. In the subsequent login, the user only needs to provide the file in the folder.

Table

Algorithm 2: Identity Authentication

INPUT: certificate/wallet , mspid OUTPUT: result of authentication if certificate != null then wallet = new wallet() X509certificate = readX509Certificate(cetificate) identity = newX509Identity(mspid, X509certificate) wallet.put(identity) builder = Identity(wallet) if builder != null then return result of authentication end if end if if wallet != null then builder = Identity(wallet) if builder != null return result of authentication end if end if

Device Binding

Algorithm 3 is the binding pseudo-code, which is used by the edge control node to bind its own serial number and the projector’s serial number to the blockchain. When binding a device, it is necessary to determine whether the device already has a binding relationship. After a new device is successfully registered, it needs to subscribe to the MQTT topic of the device.

Table

Algorithm 3: Device Binding
INPUT: projectorID, controlBoxID OUTPUT: bind result if projectorID != null and controlBoxID !=null then return please input a couple of device end if if exist(projectorID and controlBoxID) then return the two devices are bound else bind(projectorID , controlBoxID) if notExist(projectorID) then add controlBox and subscribe its MQTT topic return bind result end if end if

Resource Verify

Algorithm 4 is the resource verification pseudo-code, which is used to determine whether the received resource is consistent with the target resource. It first determines whether the projector memory is sufficient based on the file size, then determines the file type, and finally compares with the target file based on the hash value generated by MD5 algorithm.

Table

Algorithm 4: Resource Verify
INPUT: fileName, fileSize,fileType,identification OUTPUT: result of resource verify if isCacheAbleFileSize then if isCacheAbleFileType then key = MD5(fileName,fileSize,fileType) result = compare(key, identification) if result == success then if !filesExists then return verify success end if end if end if end if

Setup and Evaluation

Experiment Setup

The platform implementation is shown in Figure 7. We use a computer carrying a 2.50 GHz Intel(R) Core(TM) i5-7300HQ CPU and 16GB of RAM, in which we deploy a Hyperledger Fabric 2.3 network based on docker 20.10.21. The consensus mechanism of the blockchain employs raft consensus, which guarantees the network’s ability to tolerate faults. The block-out rule is that every 50 transactions generate a block, or generates a block in 2s. The edge node integrates the gateway program required to access the blockchain network. The app is installed in the projector. The router is used to transmit network signals, facilitating devices to be connected to the same network.

Figure 7. A security-enhanced advertising platform.

Multi-User Authentification Test

The test is completed with JMeter. The user connects to the blockchain network through a client node. When connecting to the network, the user needs to provide a certificate. The CA nodes will verify the validity of the certificate. When the user provides an incorrect certificate to connect to the network, the CA will throw an exception. When the certificate provided by the user is valid, the duration of the authentication process is related to the certificate verification speed of the blockchain network. The results of multi-user authentication test is shown in Figure 8.

Figure 8. Multi-user authentification test.

As shown in the Figure 8, the base number of user threads is set to 50, and the incremental interval of users is also set to 50 seconds. As the number of authenticated users increases, the time for certificate verification becomes longer, and the authentication time is concentrated between 1.4 and 2.7 seconds. The experimental results show that the concurrent authentication time of multiple users is low and can meet the demand for fast authentication.

Blockchain Performance

Blockchain performance test is completed with Tape, which is a blockchain performance tool. TPS refers to the throughput of the blockchain network. The performance of the blockchain is shown in Figure 9.

Figure 9. Blockchain performance.

Blockchain performance is conducted by testing the advertiser’s plan contract. The number of connections to the blockchain network is initialized to 10, and in each connection, 10 clients interact with the network at the same time.The base number of transaction is 500, with an increment of 500. From Figure 9, it can be found that, with the increase of the number of the writing transactions, the TPS ranges between 240 and 310, with transaction completion times ranging from 5 to 16 seconds. The TPS of query transaction has been maintained between 330 and 500, with transaction times kept within 10 seconds. The experimental results show that the blockchain network performance is excellent.

Table 2 shows the average generation time of each block under 10 groups of different transaction numbers. From the table, it can be seen that the block generation time is in milliseconds. The block generation time is around 0.17 seconds.

Table 2. Average block generation time.

Transaction Number	Block Gen Time
500	0.207s
1000	0.191s
1500	0.182s
2000	0.168s
2500	0.178s
3000	0.161s
3500	0.162s
4000	0.167s
4500	0.159s
5000	0.162s

Plan Response Test

The test is debugged using Android Studio to measure the time required for the projector app to search for devices and obtain the plans. A total of 10 tests are conducted with an average response time of 2.3 seconds. The experimental results show that the plan responds quickly and the system performs well. The test results of plan response time are shown in Figure 10.

Figure 10. Plan response time.

Discussion

Digital advertising delivery can accurately target audiences based on geographic location, age, gender, and interests, thereby increasing the effectiveness of advertisements and reducing labor costs. Furthermore, remote delivery allows real-time monitoring of advertising effectiveness, enabling adjustments to the ad content for more targeted and flexible delivery. Traditional advertising delivery uses a centralized cloud platform to store ad files, but centralized storage means all files are stored in one place. If the cloud platform is attacked, it may lead to loss or leakage of user files. Moreover, the system’s efficiency is affected by the performance of the cloud platform, with poor performance potentially resulting in subpar ad delivery.

To address these issues, the proposed digital advertising platform incorporates blockchain technology and deploys multiple blockchain nodes to enhance system security. Additionally, the blockchain network can provide security and transparency for ad delivery data, ensuring trust between advertisers and the advertising platform while reducing the risk of fraud and false data. Considering that blockchain is not suitable for storing large files, the platform utilizes the IPFS network for off-chain storage to ensure the security of ad files. To alleviate the computational burden on the blockchain network and achieve load balancing for the platform, edge control nodes have been developed to reduce the number of devices accessing the blockchain network. Different from traditional applications using username and password login, the platform employs digital certificates issued by the blockchain network for authentication. The platform’s playback terminal use projectors and dimming glasses, which are more convenient to install and maintain compared to LED screens, making ad delivery more environmentally friendly and energy-efficient.

However, the proposed digital advertising platform also has limitations. The integration of dimming glasses, projectors, edge control nodes, and blockchain network requires collaboration among developers from various technical fields to ensure the stable operation and effective management of the system. Dimming glass, as a new type of glass material, is relatively expensive. Implementing and managing the blockchain network requires a significant amount of hardware resources and relatively high maintenance costs.

Conclusion and Future Work

Remote digital advertising is one of the important means of digital advertising. With generative artificial intelligence being widely used, this article attempts to use blockchain and edge computing technology to build a decentralized and security-enhanced advertising platform that solves data security, device authentication and issues such as credibility between users and advertisers. By using blockchain and IPFS, decentralized storage of data is achieved and data security is enhanced. We developed edge control nodes for controlling terminal devices and a projector app for receiving plans. In addition, the platform uses dimming glass as the playback medium to further improve the flexibility of advertising. Experiments show that while the blockchain network performance is good, edge nodes can download plans in time, and the platform can well ensure data security and device identity authentication.

In our future work, we will pay more attention to environmental protection and sustainable development. By leveraging artificial intelligence and big data analytics technology, we will precisely target advertising based on the interests and behavior patterns of the primary user groups at the placement locations, thereby further enhancing the effectiveness of ad promotion.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are openly available in Gitee at https://gitee.com/xxl2018141424/system.git, reference number system-2023.

Additional information

Funding

This work was supported by the 2023 Digital Learning Technology Integration and Application of Ministry of Education Engineering Research Center Innovation Fund Project, under grant [1311014].