Gbase database

What is the gbase database? What is the difference between gbase database and MySQL?

1. What is the gbase database?

GBase is a self-owned brand database product launched by NTU General Data Technology Co., Ltd., which has a high brand awareness in the domestic database market.

Official introduction: NTU General Large-Scale Distributed Parallel Database Cluster System, abbreviated as: GBase 8a MPP Cluster, is a distributed parallel database cluster with columnar storage and Shared Nothing architecture, featuring high performance, high availability, and high scalability , suitable for OLAP scenarios (query scenarios), can provide a cost-effective general-purpose computing platform for ultra-large-scale data management, and is widely used to support various data warehouse systems, BI systems, and decision-making support systems. Generally, enterprise-level big data platforms are common Architecture - lambda architecture

GBase series products include: new analytical database GBase 8a, distributed parallel database cluster GBase 8a Cluster, high-end transactional database GBase 8t, high-speed memory database GBase 8m/AltiBase, visual business intelligence GBaseBI, large directory service system GBase 8d, hard encryption Security database GBase 8s.

8a 8a Cluster 8t 8m BI 8d 8s
Analytical database Distributed Parallel Database Cluster High-end transactional database High-speed memory database Visual Business Intelligence Large directory service system Hard encrypted security database

data analysis:

8a: A new type of column storage database that supports rapid analysis of massive data.

8a cluster: Parallel database cluster based on cloud computing architecture and contribution node architecture, supports TB and PB level structured data storage query and distributed parallel expansion.

BI: Based on the idea of ​​j2ee and soa, use flex technology to realize data analysis, and realize complex business processing requirements with a small amount of programming.

Data Security:

8d: Directory service system

1. Data encryption

The GBase 8s security database is encrypted with a cipher card approved by the State Cryptography Administration. The cipher support includes key generation, key destruction, and key operation.

2. Data Integrity

GBase 8s uses the hash algorithm of the cryptographic card hardware to realize the integrity protection function of the database user data. The data of the database is stored in the disk file in the form of data pages. Each data page has a chksum field to protect the integrity of the data page. The chksum check code is generated by the hash algorithm of the password card. The database accesses the data Check the chksum check code of each data page to protect the integrity of the data page.

8s

Data transaction processing:

8m: Store data based on memory

8t: transactional database

business:

ACID properties

Indivisible, consistent changes, isolated operation, durable write.

CAP theory:

Consistence consistency All nodes access the same latest copy

Availability availability data update has high availability

Tolerance of network partition Partition Tolerance data cannot be consistent within the time limit, that is, a partition has occurred

Column-based NOSQL database

Cassandra: A database for distributed storage of structured data, developed by facebook AP type

HBase: a non-relational database for distributed storage of unstructured data, powerset develops CP type for processing massive data generated by natural language search

Hypertable: C++ implements bitable and HBase is similar

GBase 8a MPP Cluster

GBase 8a MPP Cluster database, NTU General large-scale distributed parallel database cluster system, referred to as: GBase 8a MPP Cluster, GBase 8a MPP Cluster and GBase 8a column storage database are analytical databases independently developed by NTU General. GBase 8a MPP Cluster is a distributed parallel database cluster with Shared Nothing architecture developed on the basis of GBase 8a column store database.

GBase 8a MPP Cluster adopts the shared nothing MPP federated architecture, flexible deployment, adopts a two-level deployment structure of computing storage nodes and management node dual clusters , has no single point of failure, and has good scalability. The management cluster supports the deployment of up to 64 management nodes; A single computing and storage cluster supports the deployment of more than 300 data nodes, and the entire cluster can support the deployment of more than 1,000 data nodes; the entire cluster can handle more than 15PB of data, and a single node can also handle more than 50TB of data. It has a federated architecture, large-scale parallel computing, Massive data compression, efficient storage structure, intelligent index, virtual cluster and mirroring, flexible data distribution, perfect resource management, online rapid expansion, online node replacement, high concurrency, high availability, high security, easy maintenance, efficient loading, etc. technical characteristics.

GBase 8a MPP Cluster Architecture

All components of the GBase 8a MPP Cluster product have Active-Active multi-active capabilities, and the three core components are: distributed management cluster GCWare, distributed scheduling cluster GCluster, and distributed storage computing cluster GNode. Their functions are:

  • GCluster: Responsible for SQL analysis, SQL optimization, distributed execution plan generation, and execution scheduling.
  • GCWare: It is used to share information between GCluster instances of each node (including cluster structure, node status, node resource status, etc.), and to provide operable nodes when controlling multi-copy data operations, and to control the data consistency status of each node.
  • GNode: It is the most basic storage and computing unit in GBase 8a MPP Cluster. GNode is responsible for the actual storage of cluster data on the nodes, and receives and executes the decomposed SQL execution plan from GCluster, and returns the execution result to GCluster. When data is loaded, GNode directly receives data from the cluster loading service and writes it to local storage space. When adopting the virtual cluster deployment strategy, different nodes can be physically isolated according to different service characteristics to form different VCs.
  • VC (Virtual Cluster): realize the independent management of a single business, including database objects such as database and table.
  • Free Nodes (idle nodes): GNode nodes that do not belong to any VC, usually standby machines or nodes to be expanded or replaced.

 

3. What is the difference between gbase database and Oracle/MySQL?

Both are SQL databases, both follow SQL statements and have similar syntax.
The biggest difference is that the table structure is different. Oracle is a traditional determinant. Small databases are faster, and large databases rely on indexes to improve efficiency.
gbase is distributed, and the data is not arranged in rows and columns but distributed in blocks, so the speed of small libraries is average, but the efficiency of large libraries such as terabytes is astonishing. The larger the database, the more obvious the advantages of gbase

Research: the difference between gbase and oracle

GBase 8s and Oracle are both high-end transactional databases, with the same performance, the same maturity, and widely recognized domestic products. GBase 8s can better meet the requirements of independent and controllable national security policies, and has more functions than Oracle.

l GBase 8s supports more database engines than Oracle.

Oracle is a single transaction engine, but GBase 8s is a multi-engine database. In addition to the transaction engine, GBase 8s also has a time series engine (time series), a NoSQL engine (supports processing NoSQL data and is compatible with MongoDB), and a spatial data engine (supports GIS).

l There are more solutions for high availability than Oracle.

In terms of function, Oracle's remote disaster recovery strategy has poor support for DDL statements, and Oracle does not support read operations until the latest version of the standby machine, while GBase 8s already supports read operations on the standby machine.

In terms of business, Oracle's Data Guard, Golden Gate, and RAC functions are not built-in functions of the database, and need to be purchased separately, and the construction process is complicated and management is difficult. The enterprise version of GBase 8s includes intra-city hot standby HDR, remote disaster recovery RSS and cluster SDS high-availability solutions for shared storage, no need to purchase separately, and the construction process is simple and easy to manage.

l It has more advantages in terms of grid and scalability.

GBase 8s is better than Oracle in terms of node expansion. There is no limit to the number of nodes. Adding more nodes will not cause performance degradation, and there will be no risk of triggering BUG. Moreover, the GBase 8s cluster can increase the throughput of the master node by improving the hardware configuration, and it can also have the data throughput capability of the dual-node write of the Oracle RAC cluster.

l The security policy complies with national policy requirements.

Based on the information security considerations advocated by the state, the security encryption module of GBase 8s is completely independently developed by Nanda General Company, which fully complies with the national policy requirements, while the Oracle encryption module is internationally accepted and does not meet the national information security requirements.

In addition, NTU General can provide all codes of the database for national review, while Oracle cannot meet this requirement.

l Adapted platforms have natural advantages.

GBase 8s can support the following common and domestic platforms

CPU: Support X86, SPARC, POWER, Loongson, Phytium, Sunway, Open Power.

Operating system: Windows series, Linux (REDHAT, SUSE, CENTOS, Winning Kylin, Zhongke Fangde, Puhua, Meditation, Hongqi, Yiming), UNIX (HP-UX, AIX, Solaris, SCO), Inspur Tiansuo K1 small machine, Sugon Dragon server, etc.

Storage device: support SAN and disk array, EMC, IBM, SUN, Seagate, LSI, HP, Inspur, Sugon, Hongshan, etc., support distributed file system (GPFS, etc.)

Middleware: Weblogic, Websphere, Tomcat, Jboss, TongWeb (Dongfangtong), Apusic (Kingdee), Zhongchuang, Puyuan

SQL standard: support SQL92, SQL99, SQL2003 standard.

Oracle can only support common platforms currently on the market, and cannot adapt to most domestic platforms.

4. Migrate the oracle database to gbase

Figure 1 shows the migration process. The data migration process consists of two parts.

The first part is to migrate the structure of the database, mainly including the migration of database objects;

The second part is to migrate the data into the target database.

Among them, the real data migration is completed in the second part. For data migration, it is usually necessary to suspend the application, so as to ensure the data consistency and integrity before and after the migration. Using conventional methods, if the data volume is large, a long window of application downtime is required. When migrating applications with relatively small time window requirements, third-party tools (such as CDC) can be used to migrate incremental data, which can minimize downtime and even achieve seamless migration from Oracle to GBase 8s.

Guess you like

Origin blog.csdn.net/qq_22473611/article/details/126592416