Wonderful Blog

One stop link tracking: Alibaba Cloud's end-to-end solution
primary
Recommendation

Author: In the hot summer of Yahai, when you open the take away app to buy milk tea, you find that the order failed; On the May Day holiday, when you are driving by yourself, you find that the navigation response is slow and you often miss the intersection; In the middle of the night, when you were tutoring your child's homework, you found that the GPT application was too late to respond. no

16:51 the day before yesterday
four hundred and ninety
one
Vivo product management in CICD implementation practice
primary
Recommendation

Author: vivo Internet Performance Platform Team - Chen Lingling In DevOps practice, product management is an important part, which can help the team quickly deliver high-quality and reliable software. This article will introduce that in the field of DevOps, vivo CICD is like

05/29 20:59
three hundred and ninety
zero
Technology sharing | When SpringBoot is streaming, why does an error report suddenly after normal output?
primary
Recommendation

Project background A SpringBoot project uses both Tomcat filters and Spring interceptors. Some thread variables are initialized in the filter and used in the interceptor. This project needs to call the big language model for streaming output. In the project, the author uses

11:38 the day before yesterday
seven hundred and twenty-three
zero
Have you considered the portability of running cloud on any cloud
primary
Recommendation

Cloud portability is a strategy for building scalable, resilient, cloud native applications. When it comes to cloud native *, cloud portability is usually implicitly considered. Cloud native is an application development and deployment architecture method * which can maximize the flexibility and agility of cloud computing resources

11:29 the day before yesterday
one hundred and eighty-four
zero
PieCloudDB Database Flink Connector: Let data flow
primary
Recommendation

In the face of various types of traditional databases that have been running for a long time in the customer's environment, how to design a data migration scheme elegantly, which can not only flexibly respond to various data import scenarios and multi-source heterogeneous databases, but also meet the customer's requirements for the accuracy, consistency and timeliness of data import results

10:36 the day before yesterday
two hundred and forty-two
zero
Mystery of MySQL Core Module | Issue 18 | What does a lock look like in memory*
primary
Recommendation

Table locks and row locks are carried by lock structures. How do these lock structures exist in memory? Author: Cao Shengchun, Aikesheng technical expert, author of the public account "One tree, one stream", focusing on the source code of MySQL and OceanBase. Produced by Aikesheng Open Source Community, original

10:18 the day before yesterday
six hundred and sixty-eight
one
How HttpSecurity assembles the filter chain
primary
Recommendation

Some friends asked this question. Write a simple article and chat with everyone. A SecurityFilterChain First of all, we all know that a lot of functions in Spring Security are implemented through filters, whether it is authentication, RememberMe Login, will

09:31 the day before yesterday
three hundred and seventy-eight
zero
2.1 million token free test Spark model V3.5, and personal cloud disk is available for trial!
primary
Recommendation

What can 100000 tokens do? Write an industry analysis report, translate a chapter of foreign novels, or talk about emotional topics and answer a complex question... In the big model world, "token" is a commonly used measurement unit. A token usually represents a text unit

05/30 18:22
6.1K
zero
Rainbond Joins Hands with TOPIAM to Create a New Experience of Enterprise Cloud Native Identity Control
primary
Recommendation

TOPIAM Enterprise Digital Identity Management and Control Platform is an open source IDaas/IAM platform used to manage accounts, permissions, identity authentication, application access, help integrate all identities of internal office systems, business systems and third-party SaaS systems deployed locally or in the cloud, and achieve

05/30 17:55
two hundred and sixty-five
zero
Analysis of GaussDB SQL Query Statement Execution Process
primary
Recommendation

As far as relational databases are concerned, cutting-edge SQL is of great importance. Just like a conductor of an orchestra, he guides the correct interpretation of the work and the harmony and unity of rhythm. As a new generation of relational distributed database, Huawei Cloud GaussDB has excellent technical performance and industrial competitiveness

05/30 17:19
one hundred and forty-eight
zero
Introduction to GaussDB lightweight operation and maintenance management tool
primary
Recommendation

Foreword This course will start from the architecture of the management platform, combined with the introduction of the functions and operations of the platform's instance management, instance upgrade, disaster recovery management and monitoring alarm, comprehensively cover the daily operation and maintenance, and guide you to understand and skillfully use the GaussDB operation and maintenance platform to complete the operation and maintenance work. 1、

05/30 17:18
four hundred and sixty-eight
zero
One article explains OceanBase stand-alone version: architecture introduction, deployment process, performance test, MySQL comparison, resource configuration, etc
primary
Recommendation

Introduction OceanBase stand-alone centralized cluster (i.e. stand-alone version, no longer referred to as "stand-alone centralized cluster" later) is a minimalist database architecture introduced by OceanBase, which is different from distributed cluster architecture. The stand-alone version has no multi copy and capacity expansion capabilities, and is suitable for development and testing environments and data

05/30 15:40
two hundred and ninety-nine
one
Baidu Comate is used to improve the efficiency of research and development, and Baidu's marketing service team creates "light barges" to accelerate the intelligent marketing
primary
Recommendation

Author | The Commercial Advertising Platform Department guides the ever-changing business development process, and effective marketing is crucial. Providing users with high-quality platform services and improving advertising efficiency and accuracy is the "top priority" of each marketing platform, as well as the research and development work of the bottom layer of the platform

05/30 10:51
four hundred and ninety
zero
Deeply explore the principle and practice of token bucket current limiting
primary
Recommendation

In today's Internet era, with the increasing number of users and requests, the performance and stability of the system are facing enormous challenges. As one of the important means to ensure system stability, current limiting algorithm is widely used in various services and applications. The core purpose of current limiting is to

05/30 10:18
six hundred and seventy-three
zero
Open a new generation of cloud native database Databind in Kubesphere
primary
Recommendation

Last Saturday, Cloud Native Meetup Beijing Station, jointly organized by KubeSphere Community, Databind Community and Horizon Technology, ended successfully in Beijing. This Meetup event invited SkyWalking PMC members, Qingyun technology architecture and observability team to be responsible for

05/30 09:59
five hundred and eighteen
zero
SVG stroke breakthrough? The way you don't know
primary
Recommendation

Limitations of SVG Stroke In the previous blog post "How to break through the limitations of SVG stroke", we discussed some limitations of SVG stroke. In this article, we pointed out that the stroke of SVG can only be displayed in the middle, that is, it is evenly distributed on the inside and outside of the graphic edge

05/29 15:31
1.3K
zero
Currently, they are in the heap length window. Do you still need RAG?
primary
Recommendation

Since the release of Google's Gemini 1.5 Pro, there have been many people in the industry behind the "cricket" RAG. On the one hand, Gemini's performance is really outstanding. According to the official technical report, Gemini 1.5 Pro can handle up to 100 tokens stably, which is quite

05/29 15:13
6.6K
two
Dry goods collection! Calico's BGP RouteReflector policy practice
primary
Recommendation

This article is shared from the "Calico BGP RouteReflector Strategy Practice" of Huawei Cloud Community. The author: You can make friends. Calico, a background container network component, supports multiple back-end modes, including Overlay's IPIP, Vxlan mode, and Underlay's BGP mode of pure routing

05/29 14:26
five hundred and fifteen
zero
Alibaba Cloud Ambassador 2.0 upgrade: 0 threshold sideline earns tens of thousands more every month, don't you want to try?
primary
Recommendation

In 2024, more and more people will take "making money" as their goal. Change the mobile phone case or wallpaper printed with "Money Matters", check your financial fortune from time to time, buy various financial books... Of course, the most important thing is to work hard. And outside of their jobs, people try to open

05/29 11:11
5.6K
zero
This article teaches you how to call Ascend C operator
primary
Recommendation

This article is shared by Shengteng CANN, the author of "One Text Teaches You How to Call Ascend C Operator" in Huawei Cloud Community. Ascend C is a programming language launched by CANN for operator development scenarios. It natively supports C and C++standard specifications, and has both development efficiency and running performance. Based on Ascend

05/29 11:01
five hundred and twenty-six
zero

No more

Loading failed, please refresh the page

Source Code Analysis of Bookie Storage Architecture | Get Things Technology
primary
Recommendation

1、 Brief Analysis of Pulsar Storage Architecture As a new generation of MQ middleware, Pulsar fully implements the idea of separation of storage and accounting in the design of the underlying architecture. The broker and Bookeeper are deployed independently. The former is responsible for traffic scheduling, aggregation and calculation, and the latter is responsible for data storage

KubeAI Large Model Reasoning Acceleration Practice | Get Things Technology
primary
Recommendation

1、 Background Recently, we have deployed a special reasoning cluster for large models in batch in the production environment, and successfully increased the reasoning speed of large models including 70B by 50%, significantly reduced the deployment cost, and stably applied it to the production environment. This paper is based on the one

Maximizing the Effective Throughput of LLM Serving
primary
Recommendation

Today's LLM applications have diverse delay requirements. For example, a chat robot may need a fast initial response (for example, less than 0.2 seconds), but only needs to match human reading speed in decoding speed, while code completion requires a fast end-to-end generation time to realize

The application of DDD in the evolution of public comment trading system
primary
Recommendation

This article collates from the 73rd issue of Meituan Technology Salon, Architecture Evolution and Practice Based on Domain Driven Design (DDD), which mainly introduces the core concepts and common design ideas of DDD, and introduces the evolution process of the public comment trading system in combination with DDD, and finally makes some summary and thinking

05/10 10:23
1.2W
zero
Escape HTML+CSS
primary
Recommendation

Background At present, the mainstream technology for building interactive applications is Web technology, including HTML, CSS and JavaScript. In the past 10 years, the web technology ecology has undergone tremendous changes, including endless development frameworks, such as React, Vue

Use a pig to explain the difference between On Prem, IaaS, PaaS and SaaS
primary
Recommendation

Jeff Barr, the chief preacher of Amazon Cloud Technology, is [pushing]( https://x.com/jeffbarr/status/892767314168365056 )I posted a picture, using a pig 🐷 Clarified the difference between On Parm, IaaS, PaaS and SaaS. Although it has a long history

05/09 12:34
7.7K
two
Expose the problem of online JVM memory overflow caused by FileSystem
primary
Recommendation

Author: Ye Jidong, the big data team of vivo Internet. This article mainly introduces the whole process of analyzing and solving the problem of memory overflow caused by an online memory leak caused by the FileSystem class. Memory leak definition: a memory leak that is no longer used by programs

05/08 11:32
5.4K
one
Detailed explanation of visual language model
primary
Recommendation

The visual language model can be learned from images and texts at the same time, so it can be used for multiple tasks such as visual question and answer, image description, etc. In this article, we will take you at a glance in the field of visual language models: give an overview, understand its working principle, and figure out how to find the "model" of the destiny, and how to

05/06 11:42
7.1K
zero
Application of Model Quantification and Quantification in LLM | Material Extraction Technology
primary
Recommendation

1、 Model reasoning optimization With the landing practice of models in various scenarios, model reasoning acceleration has already become an important part of AI engineering. In recent years, the large model based on the Transformer architecture has become the mainstream, and has achieved SoTA achievements in various tasks. They are training and promoting

How to write cache code gracefully
primary
Recommendation

In daily coding practice, caching is often used to solve the high concurrency problem. Caching can be said to be the best tool to solve the traffic peak. Although the group middleware team has built the cache infrastructure, which has helped us solve most of the problems, in the actual compilation

04/29 16:20
1W
zero
Thinking and Practice of B-end Front end Code Aided Generation under the Big Model | Get Things Technology
primary
Recommendation

1、 Background: Repeated work, code specification: developers always face repeated development pain points during the development of B-end front-end code. The element modules of many CRUD pages are basically similar, but they still need to be developed manually, spending time on simple element construction, reducing the development efficiency of business requirements

Short video copy extraction is so simple
primary
Recommendation

After ten years of marriage, I traveled to the West Lake to spend ten miles in the spring breeze. Spring is always comfortable, and this March, because of the ten years of marriage with his wife, is particularly different. The two people took a day off from work extravagantly. Without telling their children, they went back to the West Lake to look for a popsicle shop 13 years ago (for the time

04/22 14:50
9.9K
zero
Damn it! My WiFi can only be used normally in rainy days
primary
Recommendation

This is a very strange network problem that the author met at home when he was in college. The father of the author is an experienced network engineer. They used a complex network setting at home to connect the high-speed commercial network of his father's company through Wi Fi bridging

04/22 10:01
2.1W
zero
The Practical Application of AI Biographic Aesthetics in Taobao
primary
Recommendation

This paper introduces how to develop and apply aesthetic standards to evaluate and improve the image quality generated by AI, especially in the e-commerce field. It is mainly divided into four steps: developing aesthetic standards, training aesthetic models, applying aesthetic models, and upgrading Taobao style models. Aesthetics

04/19 16:20
5.5K
zero
The opening rate per second exceeds 90%! Performance optimization of transaction background rendering | Get Things Technology
primary
Recommendation

1、 Foreword: Experience has always been one of the key words of the Get Things Technology Department. For front-end development, improving user experience is even more important. This article starts from the background performance optimization practice of this transaction, and introduces the overall architecture and design of the application, hoping to

Basic principles of HSF
primary
Recommendation

SOA solution - HSF (High speed Service Framework) is a service framework mainly adopted by Alibaba. Its purpose is to connect different business systems as a bridge and decouple the implementation dependencies between systems. Introduction to HSF Background The main problem of monomer application is different

04/12 16:20
7.9K
two
Practice Summary | Some Exquisiteness in Front end Architecture Design
Recommendation
turn

This article summarizes the author's doubts in the "agile" development process of daily/big business promotion, and tries to think about some solutions and solutions. In the process of front-end development and practice, I combed out some simple design solutions that could ease the "headache" of several agility

04/03 16:20
2.1K
zero
C++from Forgetting to Getting Started
Recommendation
turn

This article is mainly aimed at students who have learned and understood C++, and aims to help these students wake up the memory of C++and improve their own technical reserves. If you have never touched C++before, you can also understand the language as a whole. Audience oriented This article is mainly for Zeng

04/01 16:20
9K
three
Analysis and practice of Docker container orchestration technology
primary
Recommendation

This paper comprehensively explores the core concepts, tools and advanced applications of container orchestration technology, including Docker Compose, Kubernetes and other major platforms and their advanced functions such as network and storage management, monitoring, security, etc. In addition, the article also discusses the

04/01 15:28
1W
zero
Understand the interaction mechanism and principle of Kafka consumer and broker
primary
Recommendation

01 Preface AutoMQ is committed to building the next generation cloud native Kafka system, solving many pain points of Kafka in the past, and leading Kafka to the cloud native era. As a loyal supporter of Kafka ecology in China, we will continue to bring quality to Kafka technology lovers

03/29 18:40
7.9K
zero

No more

Loading failed, please refresh the page

 Back to top
Top