Transcript G64-304-Hyper Historian_High Performance Hyper

Dave Oravetz, US
Habib Rehman, Australia
Carl Harper, UK
Andre Scherpenzeel, UK
ICONICS 2008 Worldwide Customer Summit
Boston, MA
To provide not just a ‘data
logger’, but a ‘data historian’
Provide easy integration of data
historian with other applications
Provide competitive historian for
large projects
Compete with the specialist
products
ICONICS 2008 Worldwide Customer Summit – Boston, MA, April, 2008
2
High Performance
Large tag quantities at high rates
Small ‘data value’ footprint
Efficient long duration storage
64 bit & .NET Technology
Extendable platform allowing for future
growth
ICONICS 2008 Worldwide Customer Summit – Boston, MA, April, 2008
3
Easier data access/interaction
OPC UA for ICONICS products
API for custom applications
SQL interface
Integrate into database systems
Query from any OLE-DB client
ICONICS 2008 Worldwide Customer Summit – Boston, MA, April, 2008
4
ICONICS - Hyper Historian
Competition
GE – Proficy
Rockwell – Factory Talk
OSI – Echo Data Historian
AspenTech
– Info Plus
Wonderware – Historian 9
Simatic – IT Historian
ICONICS 2008 Worldwide Customer Summit – Boston, MA, April, 2008
5
Web Configuration
Web Viewers (Native WPF)
3D Charting
Archive
Compression
SQL Query Interface
Persistence
OPC-UA (Native)
Data Collection Acceleration
ICONICS
Hyper Historian
Standard Edition









ICONICS
Hyper Historian
Enterprise Edition








(OPC In-proc option)




Remote Collectors
Redundancy
Store & Forward
Multi-CPU
ICONICS 2008 Worldwide Customer Summit – Boston, MA
6
Hyper Historian Provides
Equal or better throughput performance
Equal or better maximum tag counts
Extensive client connectivity features
OPC UA, OPC HDA, OLE-DB, Excel, SQL
Server, native API, ICONICS Native
Extensive use of recent technologies
Vista 64/2008
Server 64/.NET
ICONICS 2008 Worldwide Customer Summit – Boston, MA, April, 2008
7
Hyper Historian Provides
Extensive client connectivity features
OPC UA, OPC HDA, OLE-DB, Excel, SQL
Server, native API, ICONICS Native
Extensive use of recent technologies
Vista 64/2008
Server 64/.NET
ICONICS 2008 Worldwide Customer Summit – Boston, MA, April, 2008
8
Comparison
ICONICS 2008 Worldwide Customer Summit – Boston, MA, April, 2008
9
Efficient
 High Performance
 High Compression
Reliable
 Redundancy
 Store and Forward
Data Accessibility
 TrendWorX64 Viewer
 SQL Query Interface
Data Editing
Easy to Configure
 Remote Configure-ability
 Web-based
 Online
Flexible, Automated Archiving
Scalable
 Flexible Architecture
10
•
Hyper-Historian Components
Data Collector
Data Logger
Workbench Configuration
3D Chart Control
SQL Query Interface
12
BizViz Solutions
• Reports
• Productivity Analytics
• Enterprise Bridging
• Custom Portals
• Mobile & Wireless
ICONICS Workbench
Third Party
Solutions
Operator Controls
• TrendWorX 2D & 3D Charts
• Manually Entered Data
• Operator Comments
Custom
Applications
Standard
SQL Query
• Configuration
• Archive Management
OPC-UA
Real-time Data Logger
Auto Archiving
Historical
Data Storage
External Storage
System Configuration
& Administration
Real-time
Data
On Event
Disk
Remote Data Collectors
RAM
Store &
Forward
. ..
• OPC DA
• OPC UA
• OPC XML DA
SQL 2005
SQL 2008
Configuration
Database
Remote Data Collectors
• SNMP
• BACnet
• Databases
Store &
Forward
Collects Data From:
 OPC DA Servers
 OPC XML DA Servers
 OPC UA Servers
 GenBroker Plugins
SNMP
BACnet
Databases
Portable
 XP 64-Bit
 Vista 64-Bit
 Server 2003 64-bit
 Server 2008 64-bit
14
Local or Remote
Deadband Filtering
Advanced Filtering
 Maximum
 Minimum
 Average
 Standard Deviation
 Totalizer
 Running Maximum
 Running Minimum
 Running Average
 Moving Maximum
 Moving Minimum
 Moving Average
 Most Recent On Time
Redundant Collectors
Store and Forward
15
Logging Options
 High Performance Disk Based
 RAM Memory-Based Persistent
Data
 Raw
 Advanced Filtered
 Aggregates
Data Types
 Boolean
 Integer
 Float
 String
Filtering
 None
 Deadband
 Swinging Door
16
Conditional Logging
 UDM Trigger Based
Memory Persistent Logging
 Configurable Logging Period
 Option to Log to Disk upon Event
 Configurable Pre and Post Event Store
Aggregate Logging
 Summary Tags for High-speed playback of Summary
Portable
 XP X64
 Vista X64
 Server 2003 X64
 Server 2008 X64
Redundant Loggers
17
Online Configuration
Web Based Configuration
Based on Web Workbench Toolkit
Consistent Look-and-Feel to
TrendWorX64 Configurator
18
50,000+ Values Logged per Second
150,000+ Tags
Collection Resolution to 10 MS
Immediate playback of Logged data
No wait for Database writes
21
Twx32
1,000 Tags @ 1 s
4% (34)
Twx64
Hyper
Historian
5%
1.8%
10,000 Tags @ 1 s
34.5%
42%
22%
25,000 Tags @ 1 s
n.a
49%
46%
50,000 Tags @ 1 s
n.a.
n.a
68%
• AMD Athlon 64 X2 Dual Core, 2.86 GHz
• 8 GB Memory
• Windows Vista64 Business SP1
Twx32
Twx64
Hyper
Historian
1,000 Tags @ 1 s
217 MB
140 MB
133 MB
10,000 Tags @ 1 s
690 MB
1460 MB
857 MB
25,000 Tags @ 1 s
n.a.
7300 MB
1480 MB
50,000 Tags @ 1 s
n.a.
n.a
2630 MB
• AMD Athlon 64 X2 Dual Core, 2.86 GHz
• 8 GB Memory
• Windows Vista64 Business SP1
Hyper Historian
Sample
Benchmark
Results
Tags
Rate
10,000
200 ms
CPU
Load
29%
25,000
500 ms
56%
50,000
1s
60%
100,000
10 s
8%
150,000
10 s
10%
Deadband Filter
Swinging Door Algorithm
“Advanced” Filtering
Small Data “Footprint”
Low overhead Binary file store
Aggregate Tags
25
102 Raw Values (per cycle)
Deadband filter: 31% (32 values)
Swinging Door filter: 12% (12 values)
Aggressive Deadband filter: 12% (12 values)
1
2
3
4
17
18
19
20
21
16
5
15
6
14
7
8
9
10
11
12
Deadband Filter
► Stores unneeded samples (6, 7, 15, 16)
► Misses inflection samples (4, 13)
13
Swinging Door
► Higher Compression
► Better accuracy
TrendWorX32 / TrendWorX64
~28 Bytes / 32-bit Float
100 Tags @ 1s for 1 day  ~300 Mbytes
Hyper Historian
~7 Bytes / 32-bit Float
100 Tags @ 1s for 1 day  ~70 Mbytes
Advanced file maintenance and Archiving
Control of Disk Space Use
Backup of Data
Easy Re-attachment of Backup
Archiving Based on
UDM Trigger
File Size
Age of Data
29
TrendWorX64 Chart Control
SQL Query Interface
ReportWorX
BridgeWorX
Productivity Analytics
Third Party
32
3D Charting Capabilities
Real-time And Historical Trend Data
Rotate to Change Perspective
Out-of-the-box Chart Categories
Plug-in New Chart Types
Combine Charts for Analysis
33
Summary View
Paging (Fwd & Back)
Zoom
34
Look and Feel of SQL Server
Various Output formats
Select, Insert, Update, Delete Support
35
Advanced Data Query Support
Time-in-State
Best Fit
Time-Weighted Average
Integral
Slope
Counter
36
API to Edit Logged Data
Edited data is Marked
Original data is kept
HDA Playback Options
Original data only
Original data with edited data
Combination – showing both
GUI for Manual Edits
API for bulk edits
37
Integrated into MonitorWorX
TraceWorX Support
GenEvents Support
38
Efficient
 High Performance
 High Compression
Reliable
 Redundancy
 Store and Forward
Data Accessibility
 TrendWorX64 Viewer
 SQL Query Interface
Data Editing
Easy to Configure
 Remote Configure-ability
 Web-based
 Online
Flexible, Automated Archiving
Scalable
 Flexible Architecture
40
• Questions ?
41
TrendWorX64 Logger
Hyper-Historian Logger
Based on TrendWorX32 Design
New Design – From Ground Up
SQL, Oracle, or MySQL database
Proprietary Database
25,000+ values / second
50,000+ Values / Second
50,000+ Tags
150,000+ Tags
Deadband Filter
Deadband & Swinging Door Filters
~28 Bytes / Logged 32-bit Float
~7 Bytes / Logged 32-bit Float
Centralized Logger
Distributed Architecture
Limited SQL Query Support
Strong SQL Query Support
…
Conditional Logging
…
Event-Based Logging
…
Data Editing / Data Insert
…
Aggregate Value Logging
Non Redundant
Redundancy / Store and Forward
HDA Server / GENESIS32 Integration
32-bit Collector
Standalone Installation for Collector
Redundancy & Store and Forward
Data Insert Grid
Data Insert Utility
Swinging Door at Collector
Optimizations (>100,000 values/sec)
MS Cluster Aware
43
Twx32 Configuration Import
SQL Query Interface Extensions
Quality Filter
SQL Server Logger
Dynamic Calculations
Ability to monitor health via SNMP
Data Collection from more sources
OPC AE
Support for more tags (>500,000)
44