Based on the design idea of ​​network system reliability, how should the corresponding test be considered and implemented? This article elaborate on the classification and content of network program reliability test.
The reliability of the network system solution mainly includes: the sustainability, maintainability, and rapid recovery mechanism of the network system. Correspondingly, the reliability testing of networking solutions can be summarized into the following categories:
1. The continuous operation capability of the network system under long-term, high-pressure, high-load, and high-frequency shock conditions, that is, DuraTIon test.
2. Evaluation of network system alarm management functions and system fault location means capabilities.
3. Automatic recovery time of the system under single point of failure.
4. Verify the combined deployment function of HA (High Availability) features such as link aggregation, MSTP, RRPP, BFD, GR, VRRP, ECMP, IRF, and adjust to achieve the best combined application effect.
5. Construct various types of attacks to attack network systems from end to end. Such tests can often be integrated in the DuraTIon test.
1 Network system continuous operation ability test
The goal of the continuous operation capability test of the network system is to pass the test of the harsher environment to ensure that the network system can still maintain stable operation under various impacts and pressures after it goes online in the customer's network environment. The test method is very clear: simulate the control plane and data plane model of the operator or industry customer network in the network test environment, keep the environment running continuously under the condition of great pressure and shock, and monitor the overall operating status of the network as a test Results data.
The design of test parameters is the key to ensure the test effect. Its content mainly includes the following parts.
1. Test the parameters of the networking equipment.
It is the most ideal environment to conduct the network test with the customer's original network model. However, considering the cost factors, it is often difficult to build a test network environment with a relatively real network 1: 1 in actual test investment. In order to realize the test purpose and fully verify the reliability of the system, the abstraction and trade-off of the test network are the key points. The principle of simplifying the abstract networking scale is to fully analyze and expose the performance pressure bottleneck of the network system, and focus on retaining the key nodes of the bottleneck in the system.
For example, for a tree network with a clear core layer structure, a common feasible solution is to use a 1: 1 network test at the core layer where the service traffic pressure is greatest. The number of devices in the core layer of the tree structure network is small, which also provides the possibility of 1: 1 networking in the test environment. The number of aggregation and access layer devices decreases layer by layer. The access layer devices use several branches to simulate real applications, and the remaining branches use multiple ports of high-performance test instruments to directly connect to the aggregation layer devices to simulate topology and traffic. The test networking can be flexibly adjusted according to the control plain and data plane path of the tested network.
2. Control plane and data plane parameters.
That is, the protocol model and the traffic model. The protocol model can get more clear data parameters through the customer's network planning and industry abstraction. There are often barriers to business understanding between application system and network system maintenance technicians, and in the network planning and design stage, the model cannot be obtained through traffic monitoring, so the traffic model is difficult to define accurately, so it has become the main factor that plagues network solution testing . A more effective method is to analyze the characteristics of the industry and combine the previous experience to set parameters of universal significance. The parameters can also be weighted according to the actual line expectations of different types of customers.
3. Oscillation coefficient and method.
Oscillation for control plane and data plane is one of the basic requirements of DuraTIon test. The network system is a dynamic environment. Various input and output changes from the edge and exit of the network system will cause greater pressure and expose more bottlenecks. And through the violent and high-frequency oscillation, a network that is worse than the actual environment of the customer is created, which can expose the defects that are deeper or need longer time to be discovered more quickly and fully. The violent winds and waves that have been set up in the test network through oscillation will give us a clearer understanding of the health and reliability of the network system.
The commonly used empirical value of the oscillation coefficient of the network system test is 30%, that is, the protocol table items and the flow rate loaded on the test system fluctuate 30% up and down periodically based on the set parameters. It can be adjusted as needed to observe the oscillation results under different frequency conditions. For example, 5 minutes, 10 minutes, 30 minutes, and 1 hour can be used as a cycle. Taking the number of routes as an example, routing oscillations cause various nodes in the entire network system to publish and delete routing information in large quantities, and trigger the migration of traffic paths, which puts greater pressure on the test system instability. During the actual test, the coefficients need to be further analyzed and adjusted flexibly according to the application characteristic network of industry customers. For example, customers in the Internet industry generally have search business bursts and have higher requirements for traffic oscillations.
The simulation of oscillation can be more conveniently achieved by commonly used test instruments in the industry, and will not be repeated in this article.
ZGAR Aurora 1200 Puffs
ZGAR electronic cigarette uses high-tech R&D, food grade disposable pod device and high-quality raw material. All package designs are Original IP. Our designer team is from Hong Kong. We have very high requirements for product quality, flavors taste and packaging design. The E-liquid is imported, materials are food grade, and assembly plant is medical-grade dust-free workshops.
Our products include disposable e-cigarettes, rechargeable e-cigarettes, rechargreable disposable vape pen, and various of flavors of cigarette cartridges. From 600puffs to 5000puffs, ZGAR bar Disposable offer high-tech R&D, E-cigarette improves battery capacity, We offer various of flavors and support customization. And printing designs can be customized. We have our own professional team and competitive quotations for any OEM or ODM works.
We supply OEM rechargeable disposable vape pen,OEM disposable electronic cigarette,ODM disposable vape pen,ODM disposable electronic cigarette,OEM/ODM vape pen e-cigarette,OEM/ODM atomizer device.
Aurora 1200 Puffs,ZGAR Aurora 1200 Puffs Pod System Vape,ZGAR Aurora 1200 Puffs Pos Systems Touch Screen,ZGAR Aurora 1200 Puffs Disposable Vape Pod System,1200Puffs Pod Vape System
Zgar International (M) SDN BHD , https://www.zgarvapor.com