COST-AWARENESS MULTI-CLOUD NETWORK FUNCTION DEPLOYMENT SYSTEM

Abstract

The invention relates to a cost-aware multi-cloud network function deployment system that enables the intelligent placement of virtualized network functions (VNFs) across multiple cloud providers, placing each network function in the most suitable cloud (public, private, or edge) by considering constraints such as cost, latency, and data sovereignty. The system offers a highly constrained optimization system for deploying VNFs across multiple clouds, considering cost, latency, reliability, and data sovereignty. The unique feature of the system is that it combines these four dimensions into a single objective function and continuously re-evaluates them using real-time pricing data.

Description

1 TARIFF COST-AWARENESS MULTI-CLOUD NETWORK FUNCTION DEPLOYMENT SYSTEM Technical Area 5 The invention enables intelligent integration of virtualized network functions (VNF) across multiple cloud providers. implementing it involves constraints such as cost, latency, and data sovereignty. by evaluating and placing each network function in the most suitable cloud (public, private, or edge) It relates to a cost-aware multi-cloud networking functionality deployment system. State of the Art Today, there are various approaches for deploying virtualized network functions (VNF). Current VNF ​​deployment approaches include manual deployment and single-cloud deployment. There are automation and resource-based deployment methods available. Manual In terms of deployment methodology, VNFs are mostly manually assigned to a specific cloud by engineers. This method cannot capture instantaneous changes in cloud prices and latency. It falls below optimum. In single-cloud automation methods, such as ETSI NFV-MANO. Orchestration frameworks automatically deploy VNFs, but typically require a single cloud provider. It operates internally; it does not perform cost / delay comparisons between providers. It only uses resources. In the deployment method based on CPU / memory utilization, most existing systems only rely on 20 It looks at it; but it doesn't consider cost, delay, reliability, and data sovereignty together. There are also some patents used in the current technology. For example, US 12,561,494 B2. Patent number [number] addresses multiple aspects of VNF deployment, including delay, capacity, and energy. It optimizes according to the target. It handles edge / cloud nodes; however, it considers different cloud providers. Collecting real-time prices via API and data sovereignty (legal territory) restriction 25 It does not include deployment features. Patent number US 10,985,963 B2, in the cloud. Automates VNF deployment; focuses on a single-provider environment, multi-cloud cost-effectiveness. It does not offer arbitrage or sovereignty restrictions. Various studies also exist in the academic literature. For example, arXiv file number 1903.11550. The study proposes affinity-based heuristic deployment in a multi-cloud environment; deployment cost is 30. and takes source proximity into account, but also real-time price collection and legal data. It does not address the limitation of sovereignty together. A study published in 2023, however, discusses delay and... Multi-objective integer programming and particle swapping that balances the number of servers It offers optimization; operates within a single data center / cloud context, reducing cross-vendor costs. It makes no difference. 35 2 There are four main shortcomings in the current systems: (1) deployment is usually with a single cloud limited, no cross-provider optimization; (2) instantaneous changes in cloud prices API (3) data sovereignty legal restrictions are not collected and reflected in the decision; not included; (4) cost, delay, reliability and sovereignty in a single objective function They are not optimized together and repositioning is not done according to post-positioning deviation. 5 In conclusion, due to the negative aspects described above and the current solutions being the subject of discussion... Due to its shortcomings, it has become necessary to make improvements in the relevant technical field. The purpose of the invention The invention represents a new breakthrough in this field, unlike the structures used in the current technology. The aim is to create a structure with different technical specifications that bring these elements together. The main goal of the invention is to integrate VNFs across multiple clouds with reduced cost, latency, reliability, and data efficiency. It offers a very limited optimization system that establishes its dominance by considering it collectively. Another aim of the invention is to demonstrate the uniqueness of the system by integrating these four dimensions into a single objective function. It combines and continuously re-evaluates with real-time price data. 15 Another objective of the invention is to enable the system that facilitates the placement of VNFs to make the placement decision. The goal is to provide the system with a mixed integer programming (MILP) model; each candidate cloud provider's current resource pricing, providers' pricing application programming The goal is to make it pull data periodically through interfaces (APIs); for each VNF request, determining the legal framework governing the data being processed and how that data is processed. The aim is to determine the geographical area in which it can be held; the cost of the deployed VNF and The goal is to ensure continuous monitoring of performance. To achieve the objectives described above, the invention involves the functionalization of virtualized network functions. (VNF) enables intelligent placement across multiple cloud providers, reducing cost. By considering constraints such as latency and data sovereignty together, each network function is optimized for at least 25%. Cost-aware multi-cloud networking that deploys to a suitable cloud (public, private, or edge) Its function is a placement system; its characteristic is: • VNF including the desired virtual network functionality and resource / latency requirements request, • Deployment 30 is the main optimization unit that decides which cloud to deploy VNF to. decision engine, • Resource profiler that determines VNF's CPU, memory, and bandwidth requirements, • Data that determines, according to legal restrictions, in which country / region the data should be stored. sovereignty controller, • The most suitable cloud that balances cost, latency, reliability, and data sovereignty constraints. 35 selects optimization solver, 3 • Cost & Latency: Gathers current pricing and latency values ​​for candidate clouds. evaluator, • A price aggregator that retrieves the real-time resource price of each cloud provider via API, • Latency meter that measures network latency and bandwidth to each cloud, • A reliability scorer that assigns a reliability rating based on the cloud's past outage records, 5 • Deployment executor that deploys VNF to the selected cloud and manages its lifecycle, • public cloud, which is a third-party public cloud provider, • private cloud, which is a private cloud located in the operator's own data center, • edge cloud node with low latency near the user It includes. 10 The structural and characteristic features and all the advantages of the invention are given in the figures below. Thanks to the detailed explanation written with references to the figures, it becomes clearer. This will be understood. Therefore, the evaluation should also take these figures and detailed explanations into account. It must be done by taking it. Ways to Help Understand the Discovery Figure 1 shows the general architecture of the system that is the subject of the invention. Explanation of Part References 1. VNF request 2. Placement decision engine 20 3. Resource profiler 4. Data sovereignty controller 5. Optimization Solvent 6. Cost & Delay Evaluator 7. Price Collector 25 8. Delay meter 9. Reliability Rater 10. Placement facilitator 11. General cloud 12. Private cloud 30 13. Edge cloud Detailed Description of Find In this detailed explanation, the preferred configurations of the invention are designed to better understand the subject matter. It is intended to facilitate understanding and will not impose any limiting effects. The invention, 35 Virtualized network functions (VNF) intelligently distributed across multiple cloud providers 4 implementing it involves constraints such as cost, latency, and data sovereignty. by evaluating and placing each network function in the most suitable cloud (public, private, or edge) It relates to a cost-aware multi-cloud networking functionality deployment system. The operating principle of the system described in the invention is as follows: When the system is activated, a placement decision is made. resource profiler (3) inside the engine (2) incoming VNF request (1) CPU, memory and band 5 The need for breadth arises and data sovereignty controllers (4) the legality of the data to be kept determines the region. The cost & delay evaluator (6) price collector (7) within each candidate pulls the instantaneous price of the cloud; latency meter (8) and reliability scorer (9) for each cloud Measures delay and reliability score. Optimization solver (5) cost, delay, sovereignty It determines the most suitable cloud by balancing reliability constraints. If the constraints are met, 10 Deployment executor (10) VNF to the selected cloud (public (11), private (12) or edge (13)) It is deployed; the performance and cost of the deployed VNF are continuously monitored, and deviations are identified. VNF is re-enabled. The originality of the system in question lies in its ability to combine these four dimensions into a single objective function, and It is the continuous reassessment based on real-time price data. 15 The system described in the invention enables VNFs to be integrated across multiple clouds with reduced cost, latency, reliability, and It offers a very limited optimization system that incorporates data sovereignty by considering it together. The system described in the invention makes the placement decision using a mixed integer programming (MILP) model. The decision variable is a binary (0 / 1) variable indicating which cloud each VNF will be assigned to. It is variable. The objective function minimizes it by weighting three terms: cloud resource cost, 20 End-to-end delay and loss of reliability (probability of failure). These three terms refer to the operator's... with weighting factors that can be adjusted according to priority (e.g., cost weight, delay weight, The reliability weight is balanced. Legal zone and capacity are added to the model as strict constraints: A VNF (Data Network) only transmits data to clouds within the legal domain to which it is subject and which have sufficient available resources. can be assigned. This formulation looks at 25 criteria (delay only or resource only) Unlike existing approaches, this is a holistic approach that considers all real-world constraints simultaneously. finds the optimum. The system described in the invention displays the current resource price of each prospective cloud provider, as well as the providers' Pricing is retrieved periodically via application programming interfaces (APIs). Thus, instead of static price assumptions, actual unit prices at the time of demand (processor-hour, 30 Memory-hour and output bandwidth charges are determined. Spot / on-demand price fluctuations also play a role. Because it was caught, the system incurs costs by placing the same VNF in different clouds at different times. They can benefit from arbitrage. The system described in the invention complies with the legal regulations governing the data it processes, for each VNF request. (e.g., KVKK or GDPR) determination and the geographical region in which the data can be stored 35 This enables its extraction. The optimization model applies this region as a strict constraint: the candidate cloud. The pool is limited to providers with data centers only in the permitted region. Therefore, a cloud that may seem attractive in terms of cost or delay is not legally compliant. They are automatically eliminated, and legal compliance is ensured by design. The system described in the invention enables continuous monitoring of the cost and performance of the installed VNF. It provides the actual cost and measured delay with estimates at the time of deployment. compares the deviation to the defined threshold (e.g., a 20% increase in cost or a delay to the target). When the value is exceeded, the optimization is run again. A new optimum is available. If the location is different, the VNF is repositioned taking into account the transportation cost; that is to say The move is only carried out if the savings from the move outweigh the cost of transportation. This is a closed system. The loop prevents an initially optimal configuration from deteriorating over time and ensures the system's 10 This enables it to constantly adapt to changes in cloud prices and load conditions.

Claims

6 REQUESTS 1. Intelligent virtualized network functions (VNF) across multiple cloud providers. implementing it involves constraints such as cost, latency, and data sovereignty. by evaluating and placing each network function in the most suitable cloud (public, private, or edge) It is a cost-aware multi-cloud network function deployment system, and its feature is; 5 • VNF including the desired virtual network functionality and resource / latency requirements request (1), • Deployment, the main optimization unit that decides which cloud to place VNF in. decision engine (2), • Resource profiler (3) that determines VNF’s CPU, memory and bandwidth requirements, 10 • Data that determines, according to legal restrictions, in which country / region the data should be stored. sovereignty controller (4), • The most suitable cloud that balances the constraints of cost, latency, reliability, and data sovereignty. select optimization solver (5), • Cost & Latency 15, which aggregates current pricing and latency values ​​for candidate clouds. evaluator (6), • Price aggregator that retrieves the instant resource price of each cloud provider via API (7), • latency meter which measures network latency and bandwidth to each cloud (8), • Reliability scorer that gives a reliability score based on the cloud’s past outage records (9), • Deployment executor 20 that deploys VNF to the selected cloud and manages its lifecycle (10), • public cloud, which is a third-party public cloud provider (11), • private cloud, which is the private cloud in the operator’s own data center (12), • edge cloud node with low latency near the user (13) It includes. 25

Images