Download method:
Cloud disk link:
Pan.baidu.com/s/1pIpfFO08…
On November 30th, 2018, Gdevops Global Agile Operations Summit ended in Guangzhou station. As the annual closing ceremony of Gdevops, Guangzhou station condensed the essence of annual operation and maintenance and database combat together with a group of technical heroes, so that everyone can harvest the technical achievements of the industry at the end of the year and look forward to the upcoming 2019 from a more cutting-edge perspective.
Playback of the essence of the main venue
Alibaba P8 expert Liu Ming —
Ali search DevOps and AIOps exploration and practice
Liu Ming, who is responsible for the construction of Ali Search Center, took Center Center as an example to carry out an in-depth interpretation of the internal architecture of DevOps basic scheduling framework, DevOps operation and maintenance management platform, and AIOps exploration and practice. When summarizing how enterprises should do a good job in AIOps, he especially emphasized that three things should be done on the basis of AIOps: 1. Implementation of efficient control — solid DevOps; 2. 2. Sampling of all operation and maintenance objects – operation and maintenance sample database; 3. Efficient iterative optimization of the algorithm — productization of the algorithm These three suggestions on infrastructure construction inspire enterprises who have no idea how to start AIOps.
Xie Tao, General Manager of Data Product Department of New Actions Network
From automation to intelligent enterprise AIOps evolution and landing
Following the topic of AIOps, Teacher Xie Tao mainly gave step-by-step implementation suggestions for the evolution and landing of AIOps in large traditional enterprises. According to his new torch network more than ten years of enterprise operational service experience, respectively from the big data, calculation, analysis, algorithm and machine learning, visualization the six parts, detailed share the new torch network covers the CMDB, intelligent monitoring, automated operations, audit, application of performance management, database SQL performance management, and other aspects of large-scale enterprise intelligence operations overall solution, It also summarizes the Roadmap of AIOps for enterprises and provides a reference for the evolution of AIOps system.
Nie Xin, Operation and maintenance Director of Tencent —
Tencent’s latest AIOps hit the ground running
With coffee operation and maintenance as the ultimate goal, Teacher Nie Xin introduced in detail their multidimensional monitoring of data and the characteristic function of DLP business life and death index from Tencent’s Cloud monitoring system. As for the practice of AIOps, he proposed five key stages: 1. Text +NLP; 2. Prediction/judgment problems; 3. Information convergence; 4. Root cause analysis 5. Root cause analysis root-Alarm enables us to have a clearer understanding of the technical implementation of the whole evolution process from traditional monitoring to big data monitoring and then to intelligent AI monitoring.
Xu Jianhui, Vice president of Database technology of Giant Sequoia
Design and practice of “computing-storage” separation architecture for distributed database
Starting from the demand for distributed database in the financial industry, Professor Xu jianhui shared several application cases of distributed database in transactional scenarios. Combined with these experiences and the development trend of NewSQL database technology, they found that “computing-storage” separation architecture would be a new mainstream architecture direction. This architecture can not only meet the requirements of database cloud, but also achieve better compatibility of SQL and distributed expansion of underlying storage through the separation of two layers. Therefore, Professor Xu Jianhui dissected the technical principle, architecture design and concrete implementation of this architecture layer by layer.
Top 10 MVP and Best Product awards of 2018
The awards ceremony
A special session was arranged in the morning of the summit. First of all, the “top 10 MVPS of 2018” was awarded to the experts/teams who shared their experiences and in-depth exchanges with the DBAPlus community through their excellent articles and keynote speeches. The community has grown and grown thanks to their support and dedication.
They are:
-
He Chun Yang, DBA team leader of FMPKINko
-
Yang Biao, senior architect of Meituan Dianping
-
Zeng Guoqin, representative of r&d group of Ele. me Big Data Platform Department
-
Guazi used car big data expert Wang Shiyang
-
JFrog Technology representative Fu Hui
-
Wen Zhengfeng, senior manager of Operation and maintenance of Xiaopeng Automobile Internet Center
-
New torch network chief architect Liang Ming tu
-
Zhang Liang, head of data research and development of JINGdong Data Department
-
Wang Lei, senior data architect of Everbright Bank
-
Xinpu Ji, representative of 360 operation and maintenance development team
Following the MVP award, there was also a heavyweight award — “Product of the Year 2018”, which honored the outstanding companies that made outstanding contributions to the industry this year, thanking them for their pioneering spirit and ingenuity to forge advanced products that drive the development of the industry.
They are:
-
Database Product of the Year 2018: PingCAP’s “New Distributed Relational Database TiDB”
-
DevOps Product of the Year 2018: JFrog Artifactory by JFrog Technology
-
Best Operation and Maintenance Product of 2018: “SIOPS Intelligent Operation and Maintenance Management Platform” of Shanghai New Ju Network Technology Co., LTD.
In the New Year, we hope that we can continue to work together with many experts, teams, enterprises and more technical fellow passengers through thick and thick. We also hope that there will be more technical forces to join our team and create more value and glory for the technology circle together!
Operation and maintenance special performance essence replay
Baidu Cloud AIOps deployment team leader Guo Yongshuai —
Baidu intelligent continuous delivery design ideas and practical experience
For Internet products, speed and quality of delivery are critical. In the face of baidu variety, size and structure of complex product status, Guo Yongshuai teachers to share their experience in intelligent continuous delivery, adhere to the “book with rail, on the same line, car homotopy”, in a unified way of operational description, operational methods and models for the principle to build the overall AIOps architecture, introduces intelligent operations continuous delivery process, make the system make decisions, Ensure quality while improving efficiency, landing intelligent continuous delivery of products.
Hui Fu, Senior Architect, JFrog
DevOps enables business Agile practice sharing
Agile DevOps is highly automated, capable of making process decisions based on key data, with predefined and highly scalable policies and rich business dimensions. Facing the challenges of identifying key data, managing delivery life cycle data, improving speed and quality based on data, and determining measurement dimensions and thresholds, Professor Fu Hui shared solutions to different challenges and recommended some practical standard models and technical architecture selection reference. He believes that to deliver improved business agility, we need to achieve a closed loop of continuous delivery, discover improvement optimization points from a business perspective, and determine the data dimension.
Guang Fa Securities operation and maintenance expert Peng Huasheng —
Operation and maintenance tool product design ideas
Teacher Peng Huasheng found a new way to share from the perspective of product design, but also brought new ideas to everyone at the scene. In the cooperative development, Party A is close to the user and Party B is close to the technology. In order to maximize the win-win situation between the two, it is necessary to complete the automation of management with the idea of scenes and visual means to promote the interconnection. The most important thing is to achieve unified visualization, complete self-service key scene construction and automation of administrative and transactional work.
360 AIOps Project leader Xinpu Ji —
360 intelligent operation and maintenance practices
At present, we all face similar problems in operation and maintenance work: explosive growth in the number of hardware and software, limited value of monitoring platform, low efficiency in fault location, and operation and maintenance personnel are often on duty 24/7. JiXinPu teacher to discuss these common problems, from two aspects of cost and efficiency consideration, choose intelligence operations road, 360 through the data accumulation, capacity forecast, classification, recycling, MySQL intelligent scheduling, KPI clustering convergence, anomaly detection, alarm and alarm events returning analysis, etc., the AI algorithm is applied to the Ops scenario, Alleviate current problems in operation and maintenance.
Wang Guansheng, Technical Director of Meitu Operation and Maintenance Department —
How does Meitu build intelligent monitoring system based on AIOps
Monitoring can help the team observe the status of products and user experience. For the monitoring operation and maintenance of meitu’s massive user data, Mr. Wang Guansheng brought their experience in building an intelligent monitoring system based on AIOps. In order to industrialize the monitoring system, it is necessary to solve the three core problems of monitoring: complete, fast and accurate, through the construction of integrated monitoring matrix to achieve end-to-end full-dimensional coverage; Build big data analysis system based on logs; A unified alarm platform (UAP) and an alarm convergence center (UDP) are set up to realize the integrated intelligent monitoring solution driven by data.
Database special replay essence
Chi Yan, P9 technical expert, Ant Financial
Operation and maintenance scenario and case practice of Ant Financial distributed database
As a financial level distributed relational database with 100% independent research and development, OceanBase is still facing challenges such as high concurrent transactions, precision and initiative, low-cost deployment, internationalization/personalization, and efficient operation and maintenance. Based on the current situation of OceanBase, Mr. Chi Yan made an in-depth analysis of the solutions to the challenges and proposed the platform development plan for the next step: vertical function extension centering on four aspects of monitoring, operation and maintenance, output and products, which can better serve users and also bring a lot of inspiration to the audience.
Li Shen, Vice President of Technology at PinCAP
TiDB architecture evolution and practical experience
As a new generation of distributed relational database, Professor Shen li said that TiDB has set clear design goals from the very beginning, ensuring the realization of horizontal expansion, high availability, ACID transactions and SQL at Scale and other functions, and formed the existing architecture on the basis of this goal. In order to better adapt to the complexity of business scenarios, TiDB will continue to evolve, improving existing functions and architectures to adapt to the changing scenarios.
Wei Wei, Technical Director of Rungan
Use data gateway to implement cross-library query
In complex business scenarios, the efficiency of data analysis is often unsatisfactory. The architecture design of database applications, data query across libraries and businesses, and data access problems caused by hierarchical data storage all affect efficiency. How to solve these problems? Teacher Wei Wei proposed to use data gateway to do cross-database query. The data gateway features decoupling, single interface, unified view, encapsulation, and complex architecture. It can adapt to scenarios such as hierarchical data storage, complex service query, and granular storage tiering, providing reference options for solutions.
Yang Biao, Senior architect of Meituan Dianping
Use case analysis of Redis
Redis can be used in many scenarios, such as caching, leaderboards, counters, distributed sessions and distributed locks, social networks, and messaging systems. Teacher Yang Biao brought common cases of Redis, including distributed lock, red envelope, Feed flow and other cases as the main scene, discussed Redis in specific situations, analyzed the difficulties in business and technical design, and hoped that everyone could gain more by drawing on one example by analogy.
Tencent senior engineer Li Yude —
MySQL kernel: from relational database to cloud database, Tencent TXSQL changes and practice
In the era of cloud database, the security and stability of traditional databases are insufficient, and the performance cannot fully meet the requirements of cloud. In order to change this situation and improve database performance, Tencent Cloud opened TXSQL project. Teacher Li Yude introduced Tencent TXSQL from relational database to cloud database and practice in detail, and analyzed the optimization details in the process of this change, providing reference ideas for the audience to optimize the use of MySQL.
The grand
I would like to express my heartfelt thanks to all the experts and teachers who took photos or didn’t take photos
Special thanks to
The success of the Gdevops Global Agile Operations Summit 2018 in Guangzhou cannot be achieved without the generous help of many partners and media. We would like to express our special thanks to the following partners and media:
At this point, 2018 Gdevops Global Agile Operations Summit has come to a successful conclusion! Thank you for your support and company in the past three years. In the future, we will make persistent efforts to focus on technical hot spots and gather various forces to provide you with higher quality content! See you next year