Use Case: Mainframe Offloading
New applications tend to have a lot of read operations and high concurrency features, especially during the promotion period, traditional host architecture performance is not easy to extend.
MIPS’s pricing system also leads to too high analytical operating costs.
Host-based application development has fewer talents, longer cycles, technical aging, etc.
Use Case: Database Modernization
Relational database is based on a single-machine architecture, it is not easy to scale horizontally, and it is difficult to support the massive data required by modern applications.
Internet of Things, next-generation CRM and other systems require flexible data models to support unstructured and semi-structured.
Tapdata Solution
- Using MongoDB as a cache read layer, replicating data from existing RDBMS such as oracle, to MongoDB to serve the new application workload.
- Real-time monitoring of data changes in existing business libraries and synchronization to MongoDB using Tapdata Replicator’s CDC technology.
- Transform the relational table into the MongoDB JSON data structure using Tapdata’s RDM technology and keep it consistent highly with the source library data.
- Developing new business on MongoDB.
Top Challenge
Heterogeneous database model matching
Relational model based on paradigm, document model based on object. Many ETL tools don’t properly support MongoDB.
Real time synchronization capacity
Bulk migration of data is relatively simple, but real-time synchronization involves the underlying logic of the database, which is generally difficult for developers to deal with.
Critical applications require enterprise-class reliability
How to provide high availability for migration?
How to ensure data consistency?
How to deal with the scene after the downtime?
The value provided by Tapdata
Real-time heterogeneity database synchronization
Low latency to copy data from Sybase to MongoDB
Replication delay < 500ms
Automatic model conversion
Automatically implement the transformation of relational structure to JSON document structure
Support 1-1, N-1, 1-N, N-N and other data relationships
Enterprise function and support
GUI driver, no code required
Support for high availability deployment without worrying about synchronization interruptions
Data verification, fault alarm and other mechanisms to ensure reliable operation on the line
Customer case
Modernization of a government medical institution database
Present situation
The agency manages dozens of public hospitals and hundreds of clinics and is responsible for the development, operation and maintenance of dozens of IT systems in these hospitals and clinics.
Systems such as CMS and EPR have had significant performance problems affecting the experience due to the continuous growth of patient data.
The current high availability and Dr architecture cannot meet the business needs and cannot be used across the center.
Existing architecture management is complex, innovation is difficult, and it is unable to support new scenarios such as mobile apps for citizens.
New mode
Build a private cloud system and provide a PaaS-based development test release platform.
Use Docker, Spring, Microservice and other technologies to meet the needs of new development and fast online.
Use a new generation database to solve data cross-center automatic disaster recovery, massive unstructured files (such as Xray, PDF, etc.), and high concurrent read and write requirements.
In the process of gradually transitioning to the new IT model, the smooth operation of the existing system is guaranteed.
Dual mode parallel architecture
