Nodal Architecture Overview Jeyant Tamby 20 Feb 2006

1
Nodal Architecture OverviewJeyant Tamby20 Feb
2006
2
Agenda

• Background
• Document Goals and Intended Audience
• System Overview
• Component Systems
• Data Interface Requirements
• Hardware Conceptual Design
• User Interface Design
• Hardware Costs
• Objective
• Methodology
• Assumptions
• Summary

3
Document Goals and Intended Audience

• Information Technology (IT) Centric view on Nodal
  Market Systems Architecture
• Provide an understanding of ERCOT IT systems and
  their interrelationships to support the Nodal
  Market.
• Base point for
• Risk and Feasibility Analysis
• HW and Data Center Plans
• Project Delivery Patterns
• Status
• Draft version
• Living document that will evolve as we move
  forward

4
System Overview

• Functions that are changing
• Infrastructure, Training, Testing
• Energy Management
• Market Management
• Day Ahead Market Management
• Real Time Market Management
• Commercial Systems
• Reporting/Archiving
• Network Model Management
• External Interfaces
• Portal
• XML
• ICCP/RTU
• Changes in one system may have impacts on other
  systems
• Effects sequencing of projects
• Impacts amount of regression testing
• Functions that do not have major changes
• Retail Transactions
• Registration

5
Network Model Management System

• Maintains the accuracy of the Transmission Model
  that underlies all grid calculations
• Changes may come from new facilities or changes
  to limits
• Critical in modeling congestion and outages and
  determining solutions

6
Energy Management Systems

• Provides information and tools for the real time
  monitoring and secure operation of the grid
• Mission critical fault tolerant and highly
  available
• Increased operator workload
• Redesign information presentation
• Increase decision support

7
Market Management Systems

• Provides tools and information for CRR auction,
  Day Ahead Markets, and Real Time Markets and
  congestion
• Mission critical fault tolerant and highly
  available
• Complexity of correlating the various data
  sources with the tools will increase
• Increased operator workload
• Redesign information presentation
• Increase decision support

8
Commercial Systems

• Generates Settlement Statements and Invoices,
  including revisions as required manages load
  profiling, metering, and renewable energy credits
• Must be able to recover within 24 hours
• Change in interval and granularity of bids will
  greatly increase amount of data

9
Information Services Enterprise Data Warehouse
10
Data Interface Requirements
11
Hardware Conceptual Design

• Three types of servers
• Server A is powerful (typically 64 cpu) machine
  that can be configured to represent multiple
  servers in multiples of 4 cpu blocks and can
  address 1 TB RAM
• Server B is a standalone server with up to 4 cpu
  and can address 32 GB RAM
• Server C is a standalone with up to 2 cpu and can
  address 4 GB RAM

12
User Interface Design

• The design of the user interface will be based on
  a task oriented structure. This will involve the
  following
• task definition/analysis
• workload analysis
• job definition
• tool requirements definition
• tool design and testing
• tool development and testing
• console operations validation and testing

13
Hardware Costs Objective

• Use a bottom up approach to come up with a total
  amount for hardware and standard software
  licenses.
• To develop a budget
• Document assumptions and reasons for later review

14
Hardware Costs Methodology

• Analyze individual functionality and determine
  what level of computation power is required using
  the current production hardware as a reference.
  Use the CELL concept.
• Document assumptions/reasoning for computation
  requirements
• Evaluate the computation requirements for the
  following environments
• Primary, Failover, ITEST, Development, MOTE, OTS,
  MOMS, Vendor
• Server A type cells currently available in the
  market are 3 times more powerful than the ERCOT
  production environment equivalent Server A cells.
  Scale the cell requirements down by 3.
• Server A cell type is a 4 cpu, 1 TB RAM unit of
  computation power (150,653.35 per cell)
• Server B cell type is a 4 cpu, 32 GB RAM unit of
  computation power (27,500 per cell)
• Server C cell type is a 2 cpu, 4 GB RAM unit of
  computation power (6,500)

15
Hardware Costs - Assumptions

• This estimate does not include Supporting
  Infrastructure
• Facilities
• Cabling
• Networking
• Storage frames
• Switches
• Tape backup
• Disk storage
• Certain functions may be partly funded by the
  PPL.
• The Server A proposed for Nodal is the equivalent
  of an HP SuperDome
• Nodal Application architecture is similar to
  Current Zonal Architecture
• The number of cells and their type for individual
  functions is a measure of the required
  computation power. It does not necessarily mean
  the server type to be procured.
• DB Software license is based on Enterprise
  Edition with RAC and partitioning. Per CPU and
  Per Named User license models are used.
• Operating System Cost is imbedded in the CELL
  cost.

16
Hardware Costs Summary ( million)

• Total 26.12 million

17
Hardware Costs Summary

• DB Software Licenses
• Work in progress. Numbers are preliminary.
• Licensing is one time, support cost is per year
  and depends on type of license purchased.
• Per Processor license is 70,000, Support is
  15,400.
• Per Named User license is 1,400, Support is
  308.
• Existing licenses have not been discounted.
• Total DB Software License 9.8 million