BAT big three for the data processing has its own characteristics, first look at Baidu: Baidu has the user search characterization of the demand data, reptiles and Aladdin access to public WEB data; Alibaba has transaction data and credit data; Tencent has a user Relational data and social data generated based on this. Moreover, the application and realization of BAT data for different forms: Baidu data from the data + third-party cooperation, the use of research and practical use of combination; Tencent data production for sale, mainly for their products ; Alibaba fancy is the circulation of data, data distribution center.
From the root causes of the data, most of the data owned by the BAT three giants are mainly generated based on B2C services. However, with the development of network technology and the popularization of digital marketing, another big part of the big data that comes along with it is that the consumers' advertisement behavior, dynamic, brand public opinions and other consumers Behavioral related data. For the collection, management, analysis and application of these data, companies such as AdMaster data application and management solutions have emerged in China. Founded in 2006, AdMaster is a SaaS company focused on technology driven data mining, analytics, management and applications that harness the power of technology and data to create business value for your customers. It is understood, AdMaster data applications are multi-source. Currently collected and analyzed data mainly includes digital advertising effectiveness evaluation and optimization data, social media marketing data, mobile advertising effectiveness evaluation and optimization data, cross-screen optimization and performance research data, brand e-commerce data and RTMP DMP data platform and advertising The main private data management platform. It also includes third-party partner docking data and proprietary CRM system data from advertisers.
The face of such massive and multi-source data analysis and management, mainly to promote the rapid development of AdMaster business is its core data application technology. Although AdMaster's technical strength can not be compared to the BAT giants, its technical advantages are also within the reach of digital marketing. According to Lu Yirei, AdMaster's senior technical director, "Currently, AdMaster's data processing is capable of receiving over 500,000 concurrent requests per second, with more than 4TB of new data per day and more than 1,000 dimensions of 100 billion records "At the same time, in order to ensure data quality and provide highly reliable, highly available, highly scalable and high performance system services, all AdMaster services have achieved single point and multiple backup of data. The system adopts a distributed cluster and is mainly divided into a WEB cluster, Storage cluster, database cluster, computing cluster four distributed clusters. Based on massive data processing needs, AdMaster independently developed the industry's most advanced ADH (Advertising Distribution Hadoop) data processing platform. ADH platform can help AdMaster quickly enter the calculation process after the completion of data collection, can provide advertisers with data support faster, the data processing speed compared with the same industry has more than doubled. At the same time, ADH platform also has the following advantages: Reduce data expansion, optimization of the merger process, the acquisition of data directly to generate the format required by the customer to improve processing speed; Built-in advertising industry algorithm, do not need to write MapReduce can calculate PV, UV And other dimensional data; optimize HBase query, designed for social data customization to improve processing performance; integrated data task scheduling system, you can automatically adjust the computing resources based on business needs; Storm integrated, optimizing Storm transmission, reducing data latency, real-time Data Computation; Spark is integrated to optimize iterative workloads and improve system reliability.
AdMaster's technology application benefits also include AdMaster's deep understanding of data applications, its ability to mine, analyze and process multi-source data, as well as its vast cluster of real-world experiences and recommendations for algorithms, classifications, clustering, NLP and other algorithms have a lot of research talent. "Lu billion Lei stressed.
In terms of data application standards, Raymond Ray also confidently stated AdMaster's openness in data applications and collaboration. To extend the reach of AdMaster's multi-sourced data, AdMaster embraces open source technologies and adds research and development investment in distributed storage and computing (including Hadoop, Spark, Storm, etc.) as well as algorithmic optimization (including NLP, machine learning, etc.). The aim is to provide high quality services to the industry, improve industry standards and improve overall data reliability and security.
No other powerful technology can transcend the power of talent and team. According to Lui Lei, AdMaster has two technology R & D centers in Beijing and Shanghai, accounting for 60% of the total number of employees. They mainly come from Internet giants such as Baidu, Ali, Tencent and Sina. They have deep insights on big data processing technologies, including rich experience in off-line, streaming and real-time distributed computing such as Hadoop.
In the future, as the potential of data value is greatly tapped and valued, the data application technologies and systems will continue to develop rapidly under the stimulation of market demand. The technology-centric companies such as AdMaster inevitably face the challenges of industry development along the way. Of course, challenges are opportunities, and they can only seize opportunities if they dare to meet the challenges. I believe, only such leaders can lead the development and prosperity of the entire industry.