Shulou(Shulou.com)06/01 Report--

Timesten architecture

The image above is the architecture diagram of Oracle Timesten 11g shown in the official documentation.

In order to make it easier to understand, let me explain the parts one by one.

First

Shared library (Shared Libraries): mainly contains some routines for implementing SQL and some related functions. The application will call this part.

DB Cache in Timesten, which contains user data, indexes, system directories, log buffers, lock tables, and temporary space.

In terms of parameter setting, it is divided into two types: PERM SIZE and TEMP SIZE. It is only mentioned here, which will be described in detail in the later configuration.

Adminstrative programs: users, scripts, or applications explicitly call utilities to perform services such as interactive SQL, bulk copy, backup / restore, datastore migration, and system monitoring.

System processes (Database processes): background processes provide services for startup, shutdown, and application fault detection at the system level, and for loading, checkpointing, and deadlock handling at the datastore level.

Checkpoint file (Checkpoint Files): this is not quite the same as ckpt in Oracle, it is a local physical file. It writes changes in Timesten memory and log changes to the local disk, namely the checkpoint file and the log file (Log Files). When the backup is restored, Timesten will read these two files into memory for recovery.

The key process of Cache Agent:Cache Connect to Oracle. Is an intermediate process for synchronizing Oracle and Timesten information.

After introducing these more important components, you must ask, why is it faster than Oracle? If you make Oracle's DB Cache very big, will it be the same as Timesten?

Timesten is designed in accordance with the development idea of IMDB from the beginning of its design, so its query optimization, index and algorithm are all designed for IMDB. All of this will greatly reduce the overhead of Timesten and thus increase the speed. Of course, the data is in real memory, not cached. This is also one of the reasons for being fast. Compared with Oracle's DB Cache, Timesten is relatively simple and fast to locate a data. Oracle has spent a lot of time in the process of parsing SQL to Hash location memory address and then going to memory to extract data, while Timesten does not have to be so complex.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.