Dec 13, 2022 - Beijing, China - Beijing Flywheel Data Technology Co., Ltd just officially launched SelectDB Cloud – a new cloud data warehouse designed to help organizations boost their data analytics effectiveness. It is a cloud native real-time data warehouse based on the open resource database Apache Doris, and features easy-to-use, high-performance and single unified. SelectDB Cloud is available on multiple clouds and till now it has been launched on AWS, Alibaba Cloud, Tencent Cloud, and Huawei Cloud.

WHAT IS NEW DATA WAREHOUSE?

As the big data era, all kinds of data are growing at a rapid speed. Accordingly, cloud data warehouse has shown three major trends. That is, real-time, unified and cloud native. SelectDB Cloud came into being under this background, and generated by the latest technology of data warehouse. At this launch event, a group of geeks took two hours to present people with a technical feast of their innovative research and development achievements. They did give users a glimpse of what to expect from a cloud native data warehouse in the application scenarios of data analysis. Let people also believe that they should make use of modern and easy-to-use products to bring a new experience to their digitization and benefit from many practical product strategies of SelectDB.

HOW IS SELECTDB CLOUD DIFFERENT FROM OTHERS?

1.SelectDB Cloud is high-performance

Not long before the launch event, SelectDB made outstanding achievements in ClickBench, a famous global benchmark for analytical databases. Therefore, the launch event had attracted the attention of many insiders, who put great expectation on SelectDB Cloud. As expected, SelectDB Cloud was extremely cost-effective by its improvements of performance. The two main application scenarios of data warehouse are wide table aggregation scenario and multi table association scenario. SelectDB Cloud shows huge performance advantages in both scenarios. In the wide table aggregation scenario, by SSB flat testing, the performance of SelectDB Cloud is 3.4 times that of ClickHouse, 92 times that of Presto, and 6 times that of Snowflake. In the multi-table association scenario, by TPC-H sf100 testing, the performance of SelectDB Cloud is 1.5 times that of Redshift, 49 times that of ClickHouse, and 2.5 times that of Snowflake.

 2.SelectDB Cloud is cost-effective

SelectDB Cloud was developed based on the cloud native architecture of separation between storage and computation which can also bring users ultra-low use cost of resources. As evaluated, the cost of SelectDB Cloud was only 1/2-1/5 of the cost of other privatization deployment products. With its black techs, SelectDB Cloud can take full advantage of extreme performance and low resource costs at the same time. Compared with similar products, the cost was reduced by more than 50%, but the query speed was 1.5 times faster.

 3.SelectDB Cloud is single unified

SelectDB Cloud is committed to solving the complexity and redundancy of the lake warehouse. Traditional enterprises have many problems with traditional lake warehouse, such as too many components, too many interfaces, difficult maintenance, and waste of resources due to huge system set and complex architectures. In comparison, users only need to use SelectDB Cloud to meet a variety of loads, and can also support both structured and semi-structured data analysis, as well as load isolation, greatly improving computing efficiency. The technical optimization behind this comes from three core technologies, namely, mixed load, structured/semi-structured data support and integration of lake and warehouse.

Through the improvement of mixed load technology, the performance of SelectDB Cloud is 54 times that of Hive, 12 times that of Spark. Through the improvement of structured/semi-structured data support, SelectDB Cloud can achieve 4.2 times write speed, 2.3 times query speed, and only 1/5 disk space occupation compared with ElasticSearch in typical scenarios for log storage analysis. Through the improvement of integration of lake and warehouse, the performance of SelectDB Cloud is 3-5 times that of Trino.

In addition to the above key advantages, SelectDB Cloud is compatible with MySQL connection protocol, has safe and convenient connection methods, rich data import methods, hierarchical user permission system, simple and convenient management console for administrators, as well as open source, multi-cloud and other features, which can well meet the needs of users in many industries, especially those in traditional industries.

SHOULD YOU ADOPT SELECTDB CLOUD?

There were four highly customized solutions for core data analysis scenarios proposed by SelectDB at the launch event.

1.Modern data platform solution for company operation

At present, enterprises generally use the typical Lakehouse parallel architecture solution. The main problems of this solution are high complexity, low performance, and insufficient openness. SelectDB Cloud is adopted to unify the data warehouse and data lake to a single platform. It is characterized by taking cloud data warehouse as the center, providing support for multiple analysis loads. SelectDB Cloud can greatly reduce the complexity and management cost through the simple and convenient management console, and can greatly reduce the resource costs through the storage and computing separation architecture. The multi computing clusters are designed to achieve computing isolation and elasticity, and can achieve ultra-high performance with SelectDB core. Finally, the multi-cloud consistency of SelectDB Cloud ensures openness. SelectDB core is fully compatible with Apache Doris. SelectDB Cloud provides efficient data migration tools between SelectDB and Apache Doris, ensuring that enterprises can switch between SelectDB Cloud service and Apache Doris built independently at any time.

 2.Reporting and analysis solution for business

There are many reports and analysis scenarios of business, such as site statistical report and analysis, advertising reports and analysis. These scenarios require high concurrency, low latency, low data stream latency, no loss or duplication and supporting data update. SelectDB Cloud can achieve millisecond-level low latency responses for tens of thousands of concurrent reports through various technical optimization measures. For the demand of low data stream latency, the data visibility of SelectDB Cloud can reach the second level, and the data is reliable without loss or duplication. In general, SelectDB Cloud achieved high concurrency of tens of thousands of QPS. The query delay is less than 99th percentile 200ms, and billions of records are added every day.

 3.User profile and behavior analysis solution

The behavior analysis scenario has three characteristics, which are continuous and rapid changes in the table structure, complex analysis, and low query latency requirements. The scene of user portrait has two characteristics which are real-time label update and rapid crowd selection. For upstream data sources, synchronize data to SelectDB Cloud for real-time behavior analysis, and achieve low-cost data storage and mining through SelectDB Cloud's hot and cold data layering capabilities. In the storage layer, the behavior data is extracted and the label data is stored hierarchically by SelectDB Cloud. In the query layer, SelectDB Cloud provides rich analysis functions to create high-performance materialized views of orthogonal bitmaps to achieve accurate duplicate removal and retention analysis. In the data management layer, it SelectDB Cloud supports the maintenance of high-performance table structure changes and the updating of wide table partial columns. Through practice, SelectDB Cloud can acheive an average delay of less than 10s and a P95 delay of about 20s for 300 billion active data under behavior analysis scenario. Based on hundreds of billions of data, SelectDB Cloud can help users to select crowds within 1 second according to 10 labels, and within 10 seconds according to 100 labels.

 4.Log storage and analysis solution.

Log storage and analysis scenarios are characterized by large data write throughput, real-time visibility, low cost, fast, supporting semi-structured and unstructured text retrieval and sorting by time. In the aspect of write optimization, SelectDB Cloud adopts client-site real-time small batch writing, the server memory save batch write and timing compaction mechanism. For saving the storage cost, SelectDB Cloud designs column storage, inverted index, high compression ratio algorithm, separation of storage and calculation, and hot and cold data layering and so on. Meanwhile, For quickly and accurately locate the matching rows, SelectDB Cloud uses technologies such as inverted index, time-series storage model and dynamic pruning Top N algorithm to optimize retrieval efficiency. With excellent capabilities, SelectDB Cloud can achieve 4.2 times write speed, 2.3 times query speed, and only 1/5 disk space occupation compared with traditional products.

If your business belongs to this category, you can consider trying SelectDB Cloud. It has low use costs, high performance and a strong customer service team. Most importantly, it is a new company and a practitioner of new technologies.

If you want to inquire about relevant products and technologies, please contact us via marketing@selectdb.com. We will arrange professional technicians to contact you as soon as possible.

Media Contact
Company Name: SelectDB
Contact Person: Rebecca Wang
Email: Send Email
Country: China
Website: http://selectdb.com