I didn't finish coding today, so the update will be delayed until later, probably around one or two in the morning. Mainly because this chapter has been quite challenging—I've been writing for two or three hours now, but it's still lacking a significant portion. Once the update is available, just reload this chapter.

Abstract: In order to reduce the delay in multi-heterogeneous network data secure transmission, a multi-heterogeneous network data secure transmission technology based on machine learning is designed. By defining the importance of data sources and data attributes, we preprocess the multi-heterogeneous network data and establish a multipath parallel transmission architecture. On this basis, machine learning methods are used for effective bandwidth estimation and parameter filtering processing. Finally, bandwidth scheduling and channel security protocol systems are established to complete the machine learning-based multi-heterogeneous network data secure transmission. Experimental results show that the machine learning-based multi-heterogeneous network data secure transmission significantly reduces data transmission delay, decreases interruption cases, and packet loss rate, meeting the design requirements of data transmission technology.

Keywords: Machine Learning; Multi-Heterogeneous Network; Data Secure Transmission; Network Data Preprocessing; Parallel Transmission Architecture

1 Introduction

Currently, with rapid advancements in communication technology, multiple networks exhibit distinct characteristics. Years of reform and innovation have brought wireless access technology's transmission rates nearer to their limits. Against this backdrop, to satisfy diverse business needs, multi-network integration is required. However, traditional integration mechanisms cannot simultaneously and efficiently utilize network transmission resources, fail to ensure efficient high-speed business transmission, and increase energy consumption during transmission, leading to interference issues. Consequently, many scholars have researched multi-network data transmission methods. In literature [1], Shi Lingling and Li Jingzhao studied a secure data transmission mechanism in heterogeneous networks, primarily adopting a security data transmission mechanism based on an optimized AES-GCM authentication encryption algorithm and an SHA-based digital signature algorithm. In literature [2], Zhou Jing and Chen Chen researched a data security model based on heterogeneous networks, which pre-encodes data and establishes a secure transmission channel for data transmission. The above methods achieve certain effectiveness but still have shortcomings. To address these shortcomings, this paper applies machine learning methods to multi-heterogeneous network data secure transmission to resolve existing issues. Experimental results demonstrate that this research's multi-heterogeneous network data secure transmission technology effectively solves current issues and holds practical significance.

2 Multi-Heterogeneous Network Data Preprocessing

In multi-heterogeneous network data secure transmission, much data is useless, so selecting relevant data sources from multi-network data for transmission is necessary to improve transmission accuracy and efficiency. During effective data source selection, the importance of relationships between data attributes is measured [3-4], capturing highly correlated data. The calculation expression is as follows: (1) In formula (1), T represents the comprehensive table of all data sources; (i, j) indicates the correlation between example source classes. Based on determining the importance of data sources, the highest correlated data table set is selected, reducing unrelated tables. After selecting important data sources, data attributes are analyzed. A data source comprises a set of data attributes, and these attribute characteristics reflect the basic information of data to be transmitted. Mainly through measuring data tuples' correlation, analyzing tuple occurrence frequency, i.e., defining through tuple data density, the data tuple density map is shown in Figure 1. In Figure 1, ε represents the neighborhood's designated radius. Following this thought process, weights are assigned to each tuple data in the above dataset [5-7], expressed as: (2) In formula (2), w(C) indicates attribute weight, w(tk) indicates core tuple number, δ indicates anomalies, w(tb) indicates edge tuple number.

3 Multipath Parallel Transmission Architecture

After completing the preprocessing, a multipath parallel transmission architecture is established, with the main contents as follows: data is pre-segmented; the communication stream segmentation is the sender's method for dividing large data blocks into different or same-sized data units [8], with size determined by the communication stream segmentation granularity, which mainly divides into the following categories: First, in packet-level service segmentation, packets are the smallest data flow constructing units, so the segmentation method's granularity is the smallest, with independent packet probabilities, which can be sent to the sender; Second, in flow-level traffic segmentation [9], specific destination addresses are encapsulated in packet headers, then packets with the same destination address are aggregated into data flows, each being independent, distinguished by a unique flow identifier. Flow-level segmentation technology effectively mitigates data distortion's impact on multipath transmission [10]. Third, in sub-flow-level traffic segmentation, data flows heading to the same destination are split into multiple sub-flows, all sharing the same destination address, to some extent solving load imbalance issues in flow segmentation algorithms. The multipath parallel transmission architecture is shown in Figure 2. Additionally, in bandwidth aggregation architecture, the scheduling algorithm determines how business transmission methods and business sub-flows are scheduled [11], ensuring orderly arrival at the receiving end as the core. Next, the data scheduling will be discussed.

4 Bandwidth Scheduling Scheme Formulation

For data transmission in multi-heterogeneous networks, as a path's bandwidth reaches a certain value, network bandwidth continually increases, stabilizing transmission performance. Allocating excessive bandwidth can reduce spectrum utilization, leading to spectrum resource waste. With increasingly tense spectrum resources, scheduling and managing each bandwidth in multipath parallel transmission can ensure transmission performance and effectively use resources. The following main steps are implemented: First, using machine learning methods for effective bandwidth estimation, reasonably estimating each sub-flow's maximally utilizable wireless bandwidth resources and meeting high throughput requirements with minimal resources is key for bandwidth scheduling. A coupled congestion control algorithm is used to jointly control each sub-flow, expressed as: (3) In formula (3), MSS represents the maximum message length constant set by protocol; RTTi, PLRi represent sub-flow path roundtrip delays and packet loss rates, respectively. Second, parameter filtering processing; due to wireless channel diversity and time-varying characteristics, link parameters and path effective bandwidth dynamically change, with errors present. To remove errors, network parameters are filtered using the Kalman filter to obtain accurate estimates. The Kalman filter is a discrete time recursive estimation algorithm, using a current time differential recursive setup according to the current status measurement, last time status, and prediction error to compute a more accurate current status output. When studying discrete control systems, linear stochastic differential equations are used as follows: (4) In formula (4), xk, xk-1 represent state parameters at times k and k-1; Ak, Bk are system parameters, represented as matrices in multi-model systems, denoting state transition and input matrices, respectively; uk represents controlled input parameters, wk indicates computational noise. Third, bandwidth scheduling, assuming a multipath connection C contains n sub-flows, each independent, occupying a separate path for data transmission; the scheduling process is as shown in Figure 3. Based on this bandwidth scheduling process, a channel security protocol is finally established to ensure multi-heterogeneous data secure transmission. The security protocol includes SSL protocol, rule establishment protocol, tunnel information protocol, etc. The SSL protocol primarily includes authentication and encryption algorithms, and all server-side data packets are encrypted via SSL to ensure communication security, while the rule establishment protocol involves connection info and message identification, with record table matching success generating sockets; forwarding ensures data info's forward and usage on VPN technology channels. OpenVPN programming is the principal method for implementing tunnel message protocols. The client sends request command messages to establish a connection with the server. After connecting, the server writes encrypted and verified data information into the tunnel information data area according to the SSL protocol, completing data exchange and transmission with the client. Channel security protocol structure is shown in Figure 4. During data transmission, the channel security protocol is followed to complete machine learning-based multi-heterogeneous network data secure transmission.

5 Experimental Comparison

To verify the effectiveness of the designed machine learning-based multi-heterogeneous network data secure transmission technology, experimental analysis was conducted; compared to the secure data transmission mechanism in heterogeneous networks from literature [1] and the data security model based on heterogeneous networks from literature [2], three systems' effectiveness were compared. The experimental dataset used in this experiment is shown in Table 1. Data collected illustrate the increasing quantity of data chosen for this experiment to better validate the three methods' effectiveness, primarily comparing transmission delays, data transmission interruption cases, and link packet loss rates among the three methods; specific details are as follows.

5.1 Transmission Delay Comparison

By comparing the transmission delay of three methods, the comparison result is shown in Figure 5. Analyzing Figure 5 reveals that in the transmission of Google public datasets, the transmission delay of all three methods is relatively small. As the amount of transmitted data increases, the data transmission delay of all three methods increases. However, upon comparison, it can be found that the data security transmission technology based on machine learning in heterogeneous networks studied in this research exhibits the smallest transmission delay, less than the traditional two methods.

5.2 Data Transmission Interruption Comparisons

The data transmission interruptions after applying three transmission technologies are compared, and the comparison results are shown in Figure 6. Figure 6 shows that the transmission technology studied in this research results in the fewest data transmission interruptions, consistently fewer than the two traditional transmission technologies in several experiments.

5.3 Link Packet Loss Rate Comparison

The study's machine learning-based heterogeneous network data security transmission technology and two traditional transmission technologies are respectively used for data transmission. The comparison result of the packet loss rate of these three methods is shown in Figure 7. Analyzing Figure 7 reveals the highest link packet loss rate of the secure data transmission mechanism in traditional heterogeneous networks, higher than the data security model based on heterogeneous networks and the transmission technology studied in this research. In summary, the machine learning-based heterogeneous network data security transmission technology of this research has less transmission delay and lower packet loss rate compared to the two traditional transmission technologies. This is because the transmission technology pre-processes the heterogeneous network data and formulates a bandwidth scheduling plan, establishing a secure transmission protocol, thereby improving the data security transmission effect in heterogeneous networks.

6 Conclusion

This article designed a machine learning-based heterogeneous network data security transmission technology, and the effectiveness of this research technology was verified through experiments. The technology can improve data transmission efficiency and reduce the data transmission packet loss rate, having significant practical application value. However, due to the limitation of research time, the study's heterogeneous network data security transmission technology still has certain shortcomings, which need further optimization in subsequent research.

Abstract: The virtualization technology ensures the stability and fluency of information usage, cloud storage technology ensures the rationality of data distribution, and information security technology guarantees the security of big data usage and browsing.

Keywords: Computer systems, Big data, Cloud storage, Virtualization.

0 Introduction

Computer software technology can process a large amount of data in a relatively short time using a certain logic for editing and analysis, propose relevant data information needed by users, perform reprocessing, and determine relevant data content that complies with user requirements for data analysis.

1 Virtualization Technology

Virtualization technology is an innovative computer software technology that can create a new virtual machine for users in a relatively short timeframe. Virtualization technology truly achieves the rational utilization of informational resources, effective configuration and mobilization of software resources, reasonable allocation and utilization of computer software resources, and ensures that during operation, computer software does not encounter freezing or slow issues due to uneven distribution of software resources. Flexibility is a significant feature of virtualization technology, allowing computation and calculation on virtualized computational elements, achieving cross-domain sharing and collaboration of computers, processing, and switching resources as needed by users to form a new resource chain. Virtualization technology mainly includes the following categories: server virtualization, Docker container technology, with emphasis on Docker container technology. Server virtualization is based on multiple-dimensional virtualization of computers, virtualizing a single physical computer into multiple virtual logical computers, establishing a virtualization tier to connect computer hardware and logical systems, achieving specific functions through decoupling and association. The so-called virtualization tier allows multiple virtualized operating systems to run on a single physical computer, which can be switched between, and these virtualized computers can share one or several unique software and hardware resources, such as memory, motherboard, graphics card in conventional computers. With appropriate operating system support, users can freely download programs and software for their own use on them (Figure 1).

If you find any errors ( broken links, non-standard content, etc.. ), Please let us know < report chapter > so we can fix it as soon as possible.