Massively Scalable, Fully Distributed Architecture

Pegasus is designed for deployment at multiple, geographically distributed Sites, with all distributed data accessible to each individual Client (security configuration permitting). Pegasus manages SCADA Objects, using Pegasus generated & managed Invariant Keys (IKeys), across geographically distributed Sites, without “collisions” in SCADA Objects or SCADA Object history. This is particularly noteworthy because the EMS SCADA Databases themselves are not designed to avoid “collisions” across Sites.

Pegasus Data Server Applications can be deployed on a single Server at a Site, or on multiple Servers at a Site without any code changes. Furthermore, each Pegasus Data Server can execute 1 Instance of each Pegasus Server Application, or multiple Instances of each Pegasus Server Application, without any code changes. All data stored in the Pegasus RDS™ is automatically Partitioned both by Time and by Data Type.

The combination of Partitioned Data, Multiple App Instances, Multiple Pegasus Servers, Multiple Server Rings, and Multiple Geographical Sites allow Pegasus to be deployed in configurations that can capture and store extremely large quantities of real-time data with very large online time windows (measured in years). Since Scalability and Multi-Site Operations were considered at every step of the design and implementation, data distributed across Servers and across Sites can be retrieved, as if all data resided in a single, unified Data Store (e.g. Database).

The best metric for any large system is its actual operation in the field. Pegasus 2003 was placed in Production in April 2004 at 1 Site at Entergy. Pegasus 2006 is currently in Production at 5 Sites, at Entergy, operating fully 7x24 on 13 Production Servers at these 5 Sites. Pegasus, in Production, has performed the following operations with an overall availability of 99.98% over a continuous 3 year period.

• Pegasus Data Collector (PDC) has performed over 4 Trillion SCADA Database Scans, with Steady-State Loading, requiring nominal 2% CPU loading per EMS Server.
• Pegasus Data Servers (PDS) are storing, managing, and serving (as of April, 2007) approximately 6 Terabytes of data, system-wide.
• Pegasus Smart Client data retrieval, for an average Data View (typically 1 day of data for several SCADA Objects), for any time period within the current 3 years of online storage, requires less than 1 second (typically less than 200 Milliseconds).
• The Pegasus Servers in each Server Ring currently store approx 360 Billion SCADA History records, for a total record count (both Production Rings) of over 700 Billion SCADA History Records. Additionally, over 28 Million System Events are available in highly partitioned SQLServer Tables on each Server Ring.
• The PDC and PDS have managed over 840 SCADA Database Schema changes, maintaining a full SCADA Database Change History, (e.g. audit trail) captured in 46 Million Metadata records.
• The PDC and PDS are currently capturing and permanently storing real-time data for approx 320,000 SCADA Objects (Analogs, Limits, Accumulators, and Status Points) system-wide, an increase of approximately 25,000 new SCADA Objects placed in service, as Pegasus continued to run 7x24, without interruption.
• The Pegasus RDS™ operated throughout Hurricane Katrina, without incident, collecting all SCADA data, system-wide, including the detection of all Status Point Exceptions (thousands of Status Point Exceptions, per minute, during the Hurricane). A Relational Database was quickly built by Nobadeer, Inc. using this data, and delivered to Entergy Services, Inc. to assist in their post Hurricane Katrina efforts.