Abstract: The HBase version of ApsaraDB for ApsaraDB has been officially launched for commercialization after 4 months of public testing. This article introduces the past and present life of ApsaraDB for HBase (with easter eggs at the end).
With the rapid development of the Internet industry in recent years, the data volume of many businesses is increasing rapidly, and the traditional RDBMS database is gradually unable to support it. In terms of data structure, the amount of unstructured data is increasing rapidly. In short, the complexity of the database has become higher and higher from the architecture construction, program invocation to post-operation and maintenance, and many small and medium-sized companies have data scales of several terabytes and several petabytes, which requires data throughput. It has also reached millions of QPS, or even tens of millions.
Origin of HBase
In this context, HBase emerged, which is a distributed, column-oriented open source database. The technology is derived from the Google paper "Bigtable: A Distributed Storage System for Structured Data" written by Fay Chang. HBase is a NoSQL database running on Hadoop. It can integrate the key/value storage mode to bring real-time query capabilities and offline processing or batch processing capabilities through MapReduce. Hbase enables users to query records in a large amount of data, and can also obtain comprehensive analysis reports from them. Some of the current larger Internet companies such as google, facebook, alibaba began to explore HBase-based database solutions earlier.
Alibaba Group has stepped into the development and construction of HBase since 2011. It is the earliest team in China to apply, research, develop and give back to open source HBase. It also gave birth to the first Committer of the HBase community in China, becoming the first HBase community to develop in China. active evangelist. Over the past few years, Alibaba has contributed hundreds of patches to the community, making positive and significant contributions to the functions, stability, and performance of many core modules, and has become one of the important forces driving the rapid development of HBase.
Cloud Trend of HBase
Although HBase itself is excellent, it is relatively complex, involving a series of knowledge such as distribution, data storage, response delay, indexing, etc. Although its API is relatively simple, in various combinations, it is necessary to play HBase well. It also requires a lot of energy in the later operation and maintenance, and in the case of self-build, the accumulation of various costs leads to low cost performance. In this context, cloud migration has gradually become a trend.
Product Advantages of ApsaraDB for HBase
The HBase version of ApsaraDB for Alibaba Cloud launched this time . Its core is based on the deep transformation of the open source community version 1.1. Many optimization strategies that Alibaba made in the earlier version will be gradually migrated to this version. Before the official commercialization of ApsaraDB for HBase, ApsaraDB for HBase has served within the Alibaba Group for more than 6 years, and has accumulated a lot of experience in performance, operation and maintenance. Currently , ApsaraDB for HBase provides the following product advantages:
Convenient operation and maintenance:
Complete the deployment within 15 minutes, monitor and warn all indicators, expand nodes, disks and modify configurations online, automatically update and upgrade the cluster, and provide online consultation with experts
high performance:
The performance is improved by more than 30%, and the performance of individual scans is improved by 100%+
High reliability:
Multi-copy mechanism, early detection and recovery of copy loss, providing 9 data reliability
High security:
Provides whitelist isolation and supports VPC
low cost:
Compared with the self-built cost, the cloud database HBase is at least 20% cheaper. After the subsequent storage of local disk, cloud disk, and OSS, the cost will be further reduced.
The technical architecture of cloud database HBase
Basic resource layer:
The bottom layer uses the architecture of ECS and local disks to ensure high performance at low cost
Basic software layer:
Using Ali-JDK and Ali-Linux, behind the source-level team maintenance
HBase kernel layer:
The same version as that of Alibaba Group, all internal performance optimizations and functional enhancements within Alibaba can be enjoyed by customers in the public cloud. These include but are not limited to: improving read and write performance, enhancing stability, reducing disk, etc.
Operation and maintenance platform:
Including monitoring and alarming, configuration management, health diagnosis, etc., functions that are easy to operate and maintain.
Usage scenarios of ApsaraDB for HBase
ApsaraDB for HBase , as the storage in the era of big data, basically solves the following three types of scenarios:
Platform class:
Storage is a product of the platform, that is, the storage of other software, such as Kylin, which is currently very good, Alibaba's internal log synchronization tool TT, and the graph component Titan. This type of storage is often the data of the platform, and sometimes has no business meaning, and can be used as the underlying storage of the platform.
User behavior class:
This class is mainly for various business systems. The user here refers not only to people, but also to things, such as the Internet of Things. In Ali, data is mainly generated by people, such as Taobao favorites, transaction data, Want Want chat records, etc. The use here is relatively straightforward, the data is directly stored in HBase, and then read. The difficulty is that it needs to support tens of millions of concurrent write access and read, and it needs to solve the problem of service quality.
Report class requirements:
For example, reports, large screens, etc., such as Alibaba's Tmall Double Eleven large screen.
Benefits of cloud database HBase commercial launch
As an excellent product, Alibaba Cloud has conducted a 4-month free public beta test before its official commercialization. A total of 1,307 enterprises have established clusters on the cloud for testing. In order to thank all parties for their suggestions on product improvement and their contributions to product development during the public beta period, Alibaba Cloud decided to increase its efforts to give back the technical dividends to all users. This cloud database HBase commercialization was launched at the initial stage (August 10-October 31). ), all product specifications will be on sale at **annual 62% discount **, and public beta users will also enjoy expert services for free.
Click here for product purchase
Original link: http://click.aliyun.com/m/28316/