If you receive errors when attempting to view this white paper, please install the latest version of
Adobe Reader.
"
Microsoft is committed to providing you with technical information on how to better integrate Microsoft and SAP solutions, reduce complexity, and have a continued focus on TCO. The SAP certified Sales and Distribution three-tier
standard application benchmark, using SQL Server 2005 and Windows Server 2003 running on HP and Intel hardware, delivered record-breaking results with 93,000 concurrent users."
Source : Microsoft
Standard Benchmark on Hewlett-Packard ProLiant Servers
Standard Benchmark is also known as :
Standard Benchmark,
Benchmark,
Benchmark Computer,
Benchmark PC Software,
Benchmark Performance,
Benchmark Test,
Benchmark Testing,
Benchmark Tests,
Benchmark Tools,
Benchmark Program,

Benchmark Software,
Benchmark Utilities,
Benchmark Utility,
Purchase Order,
Purchase Order Forms,
Purchase Order Management,
Purchase Order Number,
Purchase Order Processing,
Purchase Order Request,
Purchase Order Software,
Purchase Order SolutionsBenchmarking,
Benchmarks Standards,
Computer Benchmark Software,
Computer Benchmark Test,
Computer Benchmarking.
Results Summary
The benchmark was conducted around three core
scenarios:
Scenario 1: Mixed Workload
- Description: Scenario to showcase ability to run
mixed workloads, including sales order
processing, purchase order processing, and
ledger posting, without hitting scalability or
response time outage issues.
- Goal:Provide a good sizing data point for
customers/partners for the AOS and database
servers, with representative functionality.
Scenario 2: Complex Sales Order Processing
- Description:Scenario provides an additional data
point by processing complex sales orders alone.
This scenario excludes purchase orders and
general ledger postings, which are far lighter.
- Goal:Provide a good sizing data point for
customers around AOS and database servers,
with representative functionality.
Scenario3:Simple Sales Orders,Industry Compare
- Description:This scenario has the following
functional elements disabled to reduce the
functional footprint by approximately half: credit
limit checking, markup transactions, automatic
reservations, commission calculation, and trade
agreements.
- Goal:
- Industry benchmarks often test limited
functionality. This scenario is a
comparison point with industry benchmarks around Order to Cash /
Sales & Distribution Scenarios.
- Provide a data point on core technology
platform scalability for simple well
optimized transactions.
High Level Results Summary
| Response Time Measurement |
Scenario 1: Mixed Workload |
Scenario 2: Complex Sales Orders |
Scenario 3: Industry Compare |
| Benchmark Core Compare |
| Concurrency |
1,000 Users |
1,000 Users |
3,000 Users |
| AOS Server Count |
8 |
8 |
8 |
| Users Per AOS |
125 |
125 |
375 |
| % Utilization per AOS |
40.29% |
45.21% |
68.22% |
| Database Server Utilization |
38.26% |
44.14% |
82.02% |
| Throughput (Lines Per Hour) |
| Line Throughput Per Hour |
56,645 |
55,374 |
165,045 |
| Response Time (In Seconds) |
| Sales Order Header Creation |
0.18 |
0.20 |
0.48 |
| Sales Order Line Creation |
0.40 |
0.43 |
0.46 |
| Sales Order Packing Slip Creation |
0.66 |
0.71 |
1.22 |
| Sales Order Picking List Generation |
0.64 |
0.68 |
1.25 |
| Sales Order Invoice |
1.68 |
1.97 |
2.03 |
| Ledger Header Creation |
0.08 |
N/A |
N/A |
| Ledger Line Creation |
0.10 |
N/A |
N/A |
| Response Time Measurement |
Scenario 1: Mixed Workload |
Scenario 2 Complex Sales Orders |
Scenario 3: Industry Compare |
| Ledger Posting |
0.27 |
N/A |
N/A |
| Purchase Order Header Creation |
0.13 |
N/A |
N/A |
| Purchase Order Line Creation |
0.15 |
N/A |
N/A |
| Purchase Order Receipts List Creation |
0.32 |
N/A |
N/A |
| Purchase Order Invoice Creation |
0.74 |
N/A |
N/A |
Benchmark Transaction Profile – Scenario 1: Mixed Workload
The workload consists of the transactions described
in the following table. The ratio represents the user
breakdown in addition to the overall transaction rate
breakdown.
| Transaction Detail |
| Sales Order Processing |
[Trade Agreements]
20% Lines Receive Standard Discounts
Credit Limit Checking @ Line Level
80% of Lines from 1,000 Item Pool
20% of Lines from 5,000 Item Pool
Automatic Reservations
Markup Freight @ Order Level
[Miscellaneous Charges]
Detailed Sales Tax [6 Tax Codes]
Detailed Tax and Chart of Accounts
Update
Commission Calculation
Cost Accounting Update
Credit Limit Checking @ Invoicing |
| Purchase Order Processing |
Markup Transactions @ Order Level |
| Ledger Posting |
Create New Ledger
Add 5 Lines
Post Ledger |
Scenario Mix
|
Number of Users |
Transactions Per Hour Per User |
Lines Per Hour Per User |
| Sales Orders |
800 |
10 |
50 |
| Purchase Order |
150 |
10 |
50 |
| Ledger Posts |
50 |
10 |
50 |
Benchmark Transaction Profile – Scenario 2: Complex Sales Orders
This workload is sales orders only. It resembles
Scenario 1, but without the purchase orders or the
General Ledger postings. Sales orders are larger and
more processing-intensive transactions and are
benchmarked separately to provide an additional
data point.
| Transaction Detail |
| Sales Order Processing |
[Trade Agreements]
20% Lines Receive Standard Discounts
Credit Limit Checking @ Line Level
80% of Lines from 1,000 Item Pool
20% of Lines from 5,000 Item Pool
Automatic Reservations
Markup Freight @ Order Level
[Miscellaneous Charges]
Detailed Sales Tax [6 Tax Codes]
Detailed Tax and Chart of Accounts
Update
Commission Calculation
Cost Accounting Update
Credit Limit Checking @ Invoicing |
Scenario Mix
|
Number Of Users |
Transactions Per Hour Per User |
Lines Per Hour Per User |
| Sales Orders |
1,000 |
10 |
50 |
Benchmark Transaction Profile – Scenario 3: Simple Sales Orders for Industry Compare
This workload is sales orders only. It differs from
Scenario 2 in the following ways:
- Functionality is significantly reduced to enable
comparing Microsoft Dynamics AX 4.0 with other
industry benchmarks that test limited
functionality.
- This scenario is not suitable for sizing core sales
order functionality for production runs, unless
the functional settings are similar.
- The following functionality from Scenario 1 and 2 was disabled:
- Trade Agreements
- Credit Limit Checking
- Markup Transactions
- Automatic Reservations
- Commission Calculation
- Shipping Information for Order
|
Transaction Detail |
| Sales Order Processing |
80% of Items - 1,000 Item Pool, 20%
from 5,000 Item Pool
[Miscellaneous Charges]
Detailed Sales Tax [6 Tax Codes]
Detailed Tax and Chart of Accounts
Update [Ledger]
Cost Accounting Update |
Scenario Mix
|
Number Of Users |
Transactions per Hour Per User |
Lines Per Hour Per User |
| Sales Orders |
3,000 |
10 |
50 |
Benchmark Methodology
Microsoft Visual Studio 2005 Team System was used as the load driver, simulating concurrent users
through Microsoft Dynamics AX .NET Business
Connector. A business transaction was simulated at
an average rate of once every six minutes for each
concurrent user.
Measurements were recorded on all servers and were
measured when the concurrency reached steady
state. Steady state was maintained for a minimum of
60 minutes with repeat runs within acceptable
deviation of throughput and response times.
Notes:
- All runs had a minimum steady state of 1 hour.
- Ramp up durations were between 20 and 45
minutes depending on the scenario.
- Scenarios were considered successful when
deviations between runs were within 3% in terms
of response time and system utilization.
- A benchmark user logs on through a business
connector. Every effort has been made to
simulate standard client behavior. Some
differences exist between standard client and
business connector users.
Hardware Layout and Configuration
All scenarios were run using the same hardware
configuration of 8 homogeneous AOS servers
communicating with a single instance Microsoft SQL
Server™ 2005 database:
AOS Server Specification
Model HP DL385
Processor 2 x 2.6 gigahertz (GHz) AMD64
Memory 4 GB ECC
HardDisk (BootDrive) 2 x 72 gigabyte (GB) 15K RPM
Raid1
HardDisk (PageFile) 1 x 36 GB 15K RPM
Network HP NC7782 Gigabit Ethernet
AOS Software Settings
3 GB Switch
32-bit version of Microsoft Windows Server® 2003
with Service Pack SP1 (SP1), Enterprise Edition
Single AOS instance per server
Visual Studio Team System Client Specification
Model HP DL145
Processor 2 x 1.8 GHz AMD64
Memory 2 GB ECC
HardDisk Maxtor 40 GB IDE
Network Broadcom NetXtream Gigabit Ethernet
Visual Studio Team System Software Settings
32-bit version of Windows Server 2003 with SP1,
Standard Edition
Database Server Specification
Model HP DL585
Processor 4 x 2.6 GHz AMD64
Memory 16 GB ECC
HardDisk (BootDrive) 2 x 72 GB 15K RPM Raid1
HardDisk (PageFile) 1 x 36 GB 15K RPM
Network Broadcom NetXtream Gigabit Ethernet
Database Software Settings
3 GB Switch
64-bit version of Windows Server°2003 with SP1,
Enterprise Edition
32-bit version of Microsoft SQL Server 2005 with
Service Pack 1 (SP1)
Database Disk Configuration
EMC CX600 /w 4 GB 50/50
Raid Group 0 14 x 73 GB 10K RPM Raid 0 (Database)
Raid Group 1 04 x 73 GB 10K RPM Raid 0 (Log)
Raid Group 2 14 x 73 GB 10K RPM Raid 0 (Database)
Raid Group 3 10 x 73 GB 10K RPM Raid 0 (TempDB)
Benchmark Data Composition
The benchmark was run on a 177 GB base database.
The system was configured as follows:
- 50,000 customers across 100 customer groups
- 100,000 items across 100 item groups
- 20,000 vendors across 100 vendor groups
- Pricing enabled for all items ( Quantity >10 gets
5% line discount or total sale > $ 150 gets 2%
discount )
- Chart of Accounts set up to generate 12 to 22
entries per 5 line sales order
- 6 Sales Tax Codes set up for orders
- 10 different Miscellaneous Charges per company
- Commission calculation set up
- Shipment enabled for core scenarios
- History:
- 2.5 million invoiced sales orders, 12.5
million lines
- 100,000 purchase orders, 500,000
purchase order lines invoiced
Disclaimer
These benchmark results were returned in a
controlled lab environment, without other
applications running during execution. The
benchmark was executed on optimized hardware,
using the Microsoft Dynamics AX 4.0 SYS layer
without reporting activity during execution. This
benchmark is accurate only for the listed hardware,
non-customized version of Microsoft Dynamics AX,
transaction mix, data composition, and indexes.
Microsoft Dynamics is publishing a Benchmark
Toolkit for Microsoft Dynamics AX before the end of
2006 on PartnerSource and CustomerSource. Any
customer or partner running benchmarks on their
own system should expect varying results based on
their hardware, customizations, transaction mix, data
composition, and indexes. Transaction mix and data
composition will have an effect on sizing and
hardware requirements.
© 2006 Microsoft Corporation. All rights reserved.
Microsoft, the Microsoft Dynamics Logo, Visual Studio, Windows, and Windows Server are either registered trademarks or trademarks of Microsoft
Corporation, FRx Software Corporation, or Microsoft Business Solutions ApS in the United States and/or other countries. Microsoft Business