Exercise 1 System Design Process

System Architecture Design Strategies (select here for table of contents)
1. System Design Process	2. GIS Software Technology	3. Software Performance	4. Server Software Performance
5. GIS Data Administration	6. Network Communications	7. Platform Performance	8. Information Security
9. GIS Product Architecture	10. Performance Management	11. City of Rome	12. System Implementation

Your company has approved funding to purchase new hardware for your upcoming ArcGIS Server deployment. You are interested in purchasing a server with an entry-level ArcGIS Enterprise Advanced license. You receive hardware offers from three vendors (Dell, HP, and IBM). All the server candidates satisfy your licensed core and memory requirements, and the price for each vendor offering is the same. This list identifies the processor candidates for each of these servers.

• HP server with Xeon E3-1280v6 4 core (1 chip) 3900 MHz processor
• Dell server with Xeon Gold 5122 4 core (1 chip) 3600 MHz processor
• IBM server with Xeon E3-1280v5 4 core (1 chip) 3700 MHz processor

Use the Platform Capacity Calculator to identify the best platform selection

In this exercise, you will use the Platform Capacity Calculator (PCCalc tab) to evaluate each of these server solutions.

• Use performance metrics from your current operational environment to establish a performance baseline for your existing ArcGIS Server REST mapping services.
• Evaluate peak capacity for each of the offered bundle configurations.
• Evaluate the effect of user productivity on your peak concurrent user estimates.
• Finally, evaluate optional virtual server deployment options

Step 1: Estimate current web service display complexity

Your current GIS server production environment is supported on a single platform with a 2013 Xeon E3-1270V2 4 core (1 chip) 3500 MHz processor. Public use of your web services has been growing over the past six months, and initial throughput estimates for the new web services are expected to more than double over the next year. Current server utilization is reaching 80 percent during peak throughput loads of 37,000-display transactions per hour.

• Use the Platform Capacity Calculator located on the PCCalc tab to identify the capacity of your current GIS server.

Best Practice: You can save a copy of each Platform Capacity Calculator analysis by using the Excel Move or Copy Sheet command. Use the Copy Sheet command (check Create a Copy) to create a new PCCalc tab copy, and select Move to end. Rename the new tab, and use it to complete your analysis.

• In column A, select your current platform configuration.
• In column C, select TPH as the capacity output.
• In column E, select the Physical platform.

What is the capacity of the 2013 Xeon E3-1270V2 4 core (1 chip) 3500 MHz platform?

1. AGS REST 2D V 100%Dyn 13x7 PNG24 DB: ______to_______TPH.

2. Based on current peak throughput loads (37,000 TPH), what is the complexity of the current 100 percent dynamic REST web services? ______________________________________

Step 2: Find the throughput capacity (TPH) of the Esri hardware bundles

Your first objective is to identify the performance of each server candidate, which will identify how much work the server can do. Complete the following procedure to identify AGS REST 2D V 100%Dyn 13x7 PNG24 DB throughput capacity range for each server.

• Use the Platform Capacity Calculator located on the PCCalc tab to complete your analysis.

Best Practice: Save a copy for each answer in a separate sheet.

• In column C, select TPH as the capacity output.

• In column E, select Physical platform.

• In column A, select the platform configuration for each of the candidate platforms.

What is the throughput capacity for each of the vendor server platforms?

3. HP: Xeon E3-1280v6 4 core (1 chip) 3900 MHz,______to______TPH.

4. Dell: Xeon Gold 5122 4 core (1 chip) 3600 MHz,_____to_____TPH.

5. IBM: Xeon E3-1280v5 4 core (1 chip) 3700 MHz,_____to_____TPH.

You can save multiple copies of Excel tab 1.2 by including suffixes: 1.2(2), 1.2(3), and so on.

6. Which vendor solution supports the highest capacity (peak throughput)? ______________________________________

Step 3: Determine how many concurrent users the hardware candidates can support

The Platform Capacity Calculator was developed as a simple tool that would answer several of the most common questions that Esri customers ask.

• What can this hardware do?
• What do I need to buy to meet my capacity requirements?
• Will this server configuration meet my performance needs?

The following bonus exercises demonstrate some of the additional questions you can answer with the Platform Capacity Calculator.

You can identify peak server throughput in terms of transactions per hour (TPH) and in terms of peak concurrent users. Peak concurrent users are multiplied by user productivity (displays per minute) to identify peak throughput. Complete the following procedure to identify the range of peak concurrent users supported by each server.

• Use the Platform Capacity Calculator located on the PCCalc tab to complete your analysis.

Best Practice: Save a copy for each answer in a separate sheet.

• In column C, select Users as the capacity output.

• In column E, select Physical platform.

• In column A, select the platform configuration for each of the candidate platforms.

What is the AGS REST 2D V 100%Dyn 13x7 PNG24 DB capacity for the three server candidates?

7. HP: Xeon E3-1280v6 4 core (1 chip) 3900 MHz, ______to_______users.

8. DELL: Xeon Gold 5122 4 core (1 chip) 3600 MHz, ______to_______users.

9. IBM: Xeon E3-1280v5 4 core (1 chip) 3700 MHz, ______to_______users.

10. Which vendor solution supports the highest number of users? ___________________

Step 4: How does user productivity impact concurrent user capacity?

You have decided to purchase the Dell Xeon Gold 5122 4 core (1 chip) 3600 MHz platform solution for your server environment. You plan to use this new server to publish a variety of ArcGIS Server REST 2D services for your local public community. You are asked to assess how many concurrent users the new server configuration can support. You realize that it will depend on how the users access the web applications (user productivity). You want to provide a report that shows peak concurrent users as a function of user productivity.

• Use the Platform Capacity Calculator located on the PCCalc tab to complete your analysis.

Best Practice: Save a copy for each answer in a separate sheet.

• In column C, select Users as the capacity output.
• In column E, select Physical platform.
• In column A, select HP platform configuration.
• In column G (see graphic), provide custom productivity to complete the analysis.

What are the AGS REST 2D V 100%Dyn 13x7 PNG24 DB peak concurrent users for the following use cases?

11. User productivity is two (2) displays per minute. ______to_______users.

12. User productivity is four (4) displays per minute. ______to_______users.

13. User productivity is six (6) displays per minute. ______to_______users.

{ The default web user productivity is six displays per minute. }

14. How does user productivity affect peak concurrent users? ___________________

Best Practice: Save a copy for each answer in a separate sheet.

Step 5: What is your optimum deployment strategy if you choose to use virtual servers?

Your IT department plans to deploy ArcGIS Server in a virtual server environment using a Xeon Gold 5122 3600 MHz server as the host platform. IT options include an Xeon Gold 5122 4-core server or a Xeon Gold 5122 8-core server host configuration. You are asked to provide recommendations for deploying ArcGIS Server in a virtual server environment, and to adjust your throughput expectations based on the new virtual server deployment. You still plan to use this new server environment to publish a variety of AGS REST 2D V 100%Dyn 13x7 PNG24 DB services for your local public community.

The following best practices were distributed to the IT team responsible for preparing the server environment.

• Provide 4 GB memory per vCPU for each virtual server machine.
• Provide additional GB memory for each host server core.

Use the following procedure to evaluate the TPH for the following virtual server configurations.

• In column A, select each of the available host platform configuration options.
• In column C, select TPH as the capacity output.
• In column E, select Physical or Vserver platform.
• In column F, type 2-core/node for the vCPU configuration.

What is the AGS REST 2D V 100%Dyn 13x7 PNG24 DB capacity (TPH) for the following server configurations?

15. Physical 4-core server configuration: ______to_______TPH.

Hint: In column A, select Xeon Gold 5122 4-core (1 chip) 3600 MHz platform, and in column E, select Physical platform to generate physical server results.

16. 2x 2-core virtual server with Xeon Gold 5122 4-core host: ______to_______TPH.

Hint: In column A, select Xeon Gold 5122 4-core (1 chip) 3600 MHz platform; in column E, select Vserver; in column F, select 2-core/node; and in column H, select 2 to generate 4 Virtual server core results (2x 2-core) on the 4-core host.

17. 2x 2-core virtual server with Xeon Gold 5122 8-core host: ______to_______TPH.

Hint: In column A, select Xeon Gold 5122 8-core (2 chip) 3600 MHz platform; in column E, select Vserver; in column F, select 2-core/node; and in column H, select 2 to generate 4-core Virtual server results (2x 2-core) on the 8-core host.

18. 3x 2-core virtual server with Xeon Gold 5122 8-corehost: ______to_______TPH.

Hint: In column A, select Xeon Gold 5122 8-core (2 chip) 3600 MHz; in column E, select Vserver; in column F, select 2-core/node; and in column H, select 3 to generate 6-core Virtual server results (3x 2-core) on 8-core host.

19. Based on your current operational estimates, you must purchase sufficient servers to support heavy-complexity peak throughput loads of 74,000 TPH (two times your current throughput). What is the optimum server configuration that will satisfy your peak throughput requirements? ______________________________

Answers to Exercise 1 questions

Exercise 1: Use the Platform Capacity Calculator to identify the best platform selection (page 1-31)

• 1. AGS REST 2D V 100%Dyn 13x7 PNG24 DB: 28,000 to 54,600 TPH.
• 2. Based on current peak throughput loads (37,000 TPH), what is the complexity of the current 100 percent dynamic REST web services? Heavy-complexity service
• 3. HP: Xeon E3-1280v6 4 core (1 chip) 3900 MHz, 40,200 to 78,400 TPH.
• 4. Dell: Xeon Gold 5122 4 core (1 chip) 3600 MHz, 40,900 to 79,800 TPH.
• 5. IBM: Xeon E3-1280v5 4 core (1 chip) 3700 MHz, 37,000 to 72,100 TPH.
• 6. Which vendor solution supports the highest capacity (peak throughput)? Dell
• 7. HP: Xeon E3-1280v6 4 core (1 chip) 3900 MHz, 112 to 218 users.
• 8. DELL: Xeon Gold 5122 4 core (1 chip) 3600 MHz, 114 to 222 users.
• 9. IBM: Xeon E3-1280v5 4 core (1 chip) 3700 MHz, 103 to 200 users.
• 10. Which vendor solution supports the highest number of users? Dell
• 11. User productivity is two (2) displays per minute. 341 to 665 users.
• 12. User productivity is four (4) displays per minute. 171 to 332 users.
• 13. User productivity is six (6) displays per minute. 114 to 222 users.
• 14. How does user productivity affect peak concurrent users? Higher user productivity reduces the peak number of supported users.
• 15. Physical 4-core server configuration: 40,900 to 79,800 TPH.
• 16. 2x 2-core virtual server with Xeon Gold 5122 4-core host: 30,300 to 59,100 TPH.
• 17. 2x 2-core virtual server with Xeon Gold 5122 8-core host: 42,000 to 81,900 TPH.
• 18. 3x 2-core virtual server with Xeon Gold 5122 8-corehost: 62,300 to 121,400 TPH.
• 19. What is the optimum server configuration that will satisfy your peak throughput requirements? 3x 2-core virtual servers with Xeon Gold 5122 8-core host will satisfy heavy-complexity peak throughput loads up to 82,700 TPH with minimum ArcGIS Enterprise licensing.

Lesson 1 CPT Exercise review

• Exercise 01 completed CPT Workbook
• CPT Exercise Video: Lesson 1 Exercise overview 5:21

Page Footer

Specific license terms for this content
System Design Strategies 26th Edition - An Esri ® Technical Reference Document • 2009 (final PDF release)