Shulou(Shulou.com)06/03 Report--

In recent years, due to the development of the cloud process of the government, health care and large enterprises, the major operators of Mobile, Telecom and China Unicom continue to improve their cloud services to meet the increasingly complex business needs of the government and enterprises. With the launch of government affairs and websites of provincial, municipal and regional governments at all levels, as well as the cloud operation of key businesses of major enterprises, customers have put forward higher requirements for the stability and fluency of cloud business, and the relevant operation and maintenance management departments of operators

(for example, China Mobile's network department, network management center, China Telecom's network operation and maintenance department, Enterprise Information Department, Unicom system operation and maintenance department, etc.) face greater challenges in reducing the "failure rate" and reducing the "complaint rate".

Problems encountered:

Customers have extremely high requirements for the stability and fluency of Shanghai cloud business (such as government cloud, medical cloud, enterprise cloud, etc.), and the business environment and cloud environment are very complex. Once problems arise, they can only passively accept complaints, resulting in high "complaint rate" of operators. The fault involves many factors, such as link, network delay, retransmission, CPU, storage, SDN, memory, operating system and so on. The lack of fast and effective tools for troubleshooting can not change the passive situation of too high "complaint rate".

In addition to public cloud services, the operator Information Systems Department is also responsible for the operation and maintenance of internal private clouds, while some core businesses are still running on physical servers, which causes the Information Systems Department to face the challenge of "multi-cloud hybrid management". Virtual and real performance monitoring ". As the internal network includes many areas such as data center node interconnection area, core production area, management network area and so on, the operation and maintenance department is very eager to realize the common monitoring of "north-south flow" and "east-west flow". At the same time, it also needs "virtual network and physical network association monitoring", which can achieve real-time recording and time backtracking when something goes wrong.

To quote the director of the operation and maintenance department of a provincial operator, we need to find out the reasons for the slow business. The front door is often complained by the business department, there are grievances also do not know, obviously is the problem of the application itself, will also be attributed to our operation and maintenance work has not been done in place. We urgently need a real virtualization fault location and performance monitoring

The system will bring great value to our work.

Demand:

Need omni-directional visual performance monitoring system

For cloud managers, there is no effective management without visualization. Whether in the data center or in the public or private cloud or in the hybrid cloud, end-to-end visualization has become very important. At the same time, visualization is also a presentation of big data's analysis. if the one-to-one corresponding relationship with basic resources and business applications cannot be built, management will have no way to start. It needs to involve host, virtual machine, storage, network, service path, application, etc., and covers IPM > NPM > VPM > APM.

Need an one-click means of rapid troubleshooting

Over-reliance on experience and skills when problems occur, department coordination consumes a lot of time and manpower, can not locate the fault at the first time, lack of effective basis to quickly define the fault level and determine the responsibility of the problem, which delays the MTTR assessment indicators.

Available tools: Mingchen Zhihang Yunan Network and Virtualization performance Management system

The product adopts virtual template deployment, integrates big data analysis engine and high availability mechanism, and avoids the recovery time in the event of system failure. Using an efficient storage architecture, it can respond in real time and record historical data at the same time, and it is convenient to read the data at any time.

✓ does not need to change the existing cloud platform architecture, so it is easy to deploy and avoid implementation risk.

✓ can monitor virtual machine and virtual machine, virtual machine and physical machine, north-south data.

✓ supports SAAS mode, and tenants can directly see the real situation of their own resources and applications.

Use effect

Full-stack visual performance monitoring

Full-stack visualization, using a product to solve the problem of visualization of end-to-end performance of all resources (including application performance monitoring, infrastructure performance monitoring, network monitoring, storage monitoring) to reduce the troubleshooting time to a few minutes. Provide retrospective observation mode, direct backtracking to the time of the fault in the form of a timeline, and can do retrospective analysis for the problem points.

3. Pay attention to business health, reduce the time spent on analysis, search and department coordination when business failure, and eliminate the mistrust between cloud platform operation and maintenance team and business team through root cause analysis. Identify the blind spots of east-west monitoring in the cloud architecture, provide the original packet files, and achieve effective evidence support.

Find problems in time and find performance bottlenecks

1. Proactive performance failure prediction, like a medical examination, can report and solve problems before the end user feels the impact. Many times in time to solve the Gansu Mobile, Xiamen Telecom CPU over / low configuration, storage efficiency decline brought about by the impact of related business efficiency decline.

2. The powerful data penetration analysis ability shows the root cause of the performance decline in a certain period of time. Using root cause analysis, one button distinguishes the boundary of the fault, automatically associates the logical relationship between the virtual resource pool and the application service, and provides a direct display.

Optimize the allocation of resources and save investment

1. Optimize cloud resource usage costs according to actual resource usage, and find inefficient or zombie virtual machines.

Reduce waste and provide real evidence support for scientific tuning.

two。 Intuitively see the underconfigured hosts or virtual machines, so as to avoid application performance problems caused by insufficient resources.

Automatic discovery of service dependency panorama

The terminal does not need to install plug-ins to build service dependencies. Automatically draw the graph, in-depth analysis of the dependence between virtual machines and resources and business, so that the problem can be quickly analyzed and troubleshooting. By comparing the historical service mapping with that after the change, the key component factors of application performance degradation are found. Support asset carding across multiple clouds, service relationship carding, quickly find performance bottlenecks, and view the application performance of each service through application response time and associated data.

The first time to know the user or tenant's experience, to avoid complaints or prepare contingency plans. And quickly determine the scope of influence and reasons to identify the root cause of user application delay, and analyze whether it is client delay, network delay, storage delay, process delay, application delay, or server-side delay. free from the complex analysis and troubleshooting process.

I hope my sharing can help you. Mingchen Zhihang, your network health guardian.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.