Presentation is loading. Please wait.

Presentation is loading. Please wait.

European Organization For Nuclear Research CERN Accelerator Logging Service Overview Focus on Data Extraction for Offline Analysis Ronny Billen & Chris.

Similar presentations


Presentation on theme: "European Organization For Nuclear Research CERN Accelerator Logging Service Overview Focus on Data Extraction for Offline Analysis Ronny Billen & Chris."— Presentation transcript:

1 European Organization For Nuclear Research CERN Accelerator Logging Service Overview Focus on Data Extraction for Offline Analysis Ronny Billen & Chris Roderick 15 th March 2010

2 15-Mar-2010Forum on Interfacing to the Logging Database for Data Analysis2 Outline  Logging History  Today’s CERN Accelerator Logging Service  Architecture and Current Status  Data Extraction  Summary

3 Logging History  The LHC Logging System Sub-project launched by LHC Controls Project in Sep-2001  Original mandate Analysis, design, procurement of Logging Facilities for the future LHC Controls System  Original goals Information management for LHC performance improvement Meet INB requirements for recording beam history Make available long term statistics for management Avoid duplicate logging efforts  Scope creep over the years Cover complete CERN accelerator complex LHC Hardware Commissioning 15-Mar-2010Forum on Interfacing to the Logging Database for Data Analysis3

4 15-Mar-2010Forum on Interfacing to the Logging Database for Data Analysis4 Today’s CERN Accelerator Logging Service  Persistence of logged time-series data for the lifetime of LHC  Based on the Oracle database management system  On-line availability of all data  Unique architecture: parallel reads/writes of same data  Data extraction user interface with common functionality Selection of variables, time range Fast statistics on query result Graphical visualization of data set Data extraction to different file formats  Data extraction requirements have evolved significantly First order data manipulation (interpolation, averaging,…) Applicative interface to the data extraction methods

5 PL/SQL filtered data transfer Architecture 15-Mar-2010Forum on Interfacing to the Logging Database for Data Analysis5 LDB MDB Equipment – DAQ – PLC Equipment – DAQ – FEC ffffffffffff fff f ffff f ELEC COMMCV EAU TIM f fff f QPSPIC SU Coll CNGS Exp Cryo CIET WIC VACRad BLM BETSBIC BCTBPM FGC MSMK VAC ~20 Years filtered data 7 Days raw data Filters for data Reduction

6 PL/SQL filtered data transfer Current Status 15-Mar-2010Forum on Interfacing to the Logging Database for Data Analysis6 LDB MDB Equipment – DAQ – PLC Equipment – DAQ – FEC ffffffffffff fff f ffff f ELEC COMMCV EAU TIM f fff f QPSPIC SU Coll CNGS Exp Cryo CIET WIC VACRad BLM BETSBIC BCTBPM FGC MSMK VAC ~20 Years filtered data 7 Days raw data ~ 200’000 Signals ~ 50 data loading processes ~ 5.1 billion records per day ~ 130 GB per day  46 TB per year throughput ~ 800’000 signals ~ 300 data loading processes ~ 3.8 billion records per day ~ 105 GB per day  38 TB per year stored > 300 extraction clients 0.4  2 million extraction requests per day

7 Current Status 15-Mar-2010Forum on Interfacing to the Logging Database for Data Analysis7

8 15-Mar-2010Forum on Interfacing to the Logging Database for Data Analysis8 Data Extraction – No SQL Access  Direct database access must be avoided Not scalable across all clients  Number of connections  Security considerations  Volatile infrastructure Not secure  Badly written queries / application logic will crash the entire service! Not performant  Most programming languages provide database access  Few languages optimized to work with Oracle in a performant manner

9 15-Mar-2010Forum on Interfacing to the Logging Database for Data Analysis9 CERN Accelerator Logging Service Data Extraction – Java API Spring HTTP Remoting Custom Java Applications (currently > 30) 10g AS Spring HTTP Remoting metadata JDBC TS Data JDBC Metadata TIMBER LDB MDB

10 15-Mar-2010Forum on Interfacing to the Logging Database for Data Analysis10 Data Extraction – Java API  Java has been the strategic choice for all high-level control room applications  Our Java API to the Logging Service is available since several years Well documented  http://slwww.cern.ch/~pcrops/releaseinfo/pcropsdist/dm/logging-data-extractor- client/PRO/build/docs/api/ http://slwww.cern.ch/~pcrops/releaseinfo/pcropsdist/dm/logging-data-extractor- client/PRO/build/docs/api/ Easy to use  Sample code available Heavily used (> 30 custom applications + TIMBER) Fully optimized and instrumented, essential for us to monitor and guarantee the Service Provides secure access to databases hidden on Technical Network

11 15-Mar-2010Forum on Interfacing to the Logging Database for Data Analysis11 Data Extraction – Java API  JDBC fulfils our requirements, particularly with respect to Oracle performance, as it supports: Connection Pooling Statement Caching Bind Variables Flexible Array Fetching  3-Tier architecture has many more benefits Resource pooling (connections, statements) Database protection Database isolation, since users don’t need to care about:  Database schema  Server details and login credentials  Access to Technical Network

12 15-Mar-2010Forum on Interfacing to the Logging Database for Data Analysis12 Summary  The Logging Service is being heavily used.  Data access must be controlled & monitored to ensure continuity of the overall service.  A well documented, easy to use Java API is available for data extraction.  Care should be taken as to how data is extracted, since misuse will impact the whole service, including data writing!


Download ppt "European Organization For Nuclear Research CERN Accelerator Logging Service Overview Focus on Data Extraction for Offline Analysis Ronny Billen & Chris."

Similar presentations


Ads by Google