Difference between revisions of "Login VSI Workloads"

From Login VSI Documentation
Jump to: navigation, search
 
(45 intermediate revisions by 4 users not shown)
Line 1: Line 1:
−
This article will describe the differences in the workloads of 4.1 in comparison to 4.0. With the release of Login VSI 4.1 we included 4 new workloads. Three of these workloads are improved versions of the old Medium, Light and Heavy workloads. And the last one is a newly created workload.  
+
This section describes the default workloads that come with the Login VSI product. These new and improved workloads feature a new timer mechanism, which is executed multiple times within each workload, and determines the performance of the system during testing.
  
−
==All new workloads:==
+
'''Recommended:''' [https://www.loginvsi.com/blog/login-vsi/665-simulating-vdi-users-introduction-to-login-vsi-workloads Simulating VDI Users - Introduction to Login VSI Workloads]
  
−
Moving to Login VSI 4.1 a new timer mechanism has been introduced. The timer is executed multiple times within each workload and determines the performance of the system during the test. Within the workload the different timers are easy to identify.  
+
'''Note''': For more information on the additional workloads that are available to download, please see the [https://www.loginvsi.com/documentation/index.php?title=Login_VSI_Workload_Add-ons Workload Add-ons] section.
  
−
*VSI_Timer: Login VSI 4.0 timer
+
==Task Worker==
−
*VSI_Timer41: Login VSI 4.1 timer
+
This is a light workload, which runs fewer application and starts/stops them less frequently than the other workloads, resulting in lower CPU, RAM and IO usage.  
  
−
==List of old and new Workloads==
+
The Task Worker workload uses the following applications:
−
*Light - Task worker
 
−
*Medium - Knowledge worker
 
−
*Heavy - Power worker
 
−
*1(v)CPU - Office worker
 
  
−
==From Light to Task worker==
+
*Adobe Reader
 +
*Internet Explorer
 +
*MS Excel
 +
*MS Outlook
 +
*MS Word
  
−
The Light workload was focussed mostly around Excel actions and Internet explorer actions. With the new Light workload we have made the workload more diverse and not focussed on one or two applications. We have included the following applications to the Office worker workload:
+
It also uses native Windows apps (Notepad and 7-Zip) to complete the print and zip actions used by the workload meta functions.
  
−
*Outlook
+
==Office Worker==
−
*Word
+
The main goal of this workload is to be deployed in environments that only use 1(v)CPU. It is based on the Knowledge Worker workload, but has less resource usage in comparison. However, it is more intensive than the Task Worker workload.
−
*PowerPoint
 
−
*Photo viewer
 
−
*Adobe
 
−
*Copy and zip actions
 
  
−
The actual resource usage of the workload has also changed (compared from Light to Officeworker):
+
The Office Worker workload uses the following applications:
  
−
*There is a Disk Write increase of approximately 18%
+
*Adobe Reader
−
*There is a Disk Read increase of approximately 52%
+
*Internet Explorer
−
*There is an IOPS increase of approximately 15%
+
*MS Excel
−
*There is a CPU usage increase of approximately 6%
+
*MS Outlook
 +
*MS PowerPoint
 +
*MS Word
 +
*Photo Viewer
  
−
The reason for the increase of resource usage is because we noticed that the original Light workload was too Light to be applicable as a simulated Light user. Also because the timer has changed the workload could increase its resource usage.
+
It also uses native Windows apps (Notepad and 7-Zip) to complete the print and zip actions used by the workload meta functions.
  
−
==From Medium to Knowledge worker==
+
==Knowledge Worker==
 +
The Knowledge Worker is designed for 2(v)CPU environments. This is a well-balanced intensive workload that stresses the system smoothly, resulting in higher CPU, RAM and IO usage.
  
−
Both the Medium and the Knowledge worker workloads have been designed as 2vCPU workloads. The Knowledge worker has been totally redesigned to support creating a 1vCPU workload called the office worker. The goal of the new Knowledge worker workload was to result in more stable and reliable VSImax numbers that are the same or very close to tests executed with Login VSI 4.0 Medium. The changes are:
+
The Knowledge Worker workload uses the following applications:
  
−
*Removed flash games
+
*Adobe Reader
−
*Over all balancing of the workload
+
*Freemind/Java
−
*Added copy and zip actions
+
*Internet Explorer
 +
*MS Excel
 +
*MS Outlook
 +
*MS PowerPoint
 +
*MS Word
 +
*Photo Viewer
  
−
The goal of the new Knowledge worker workload was to be more stable, more smooth and still generate the same amount of VSImax as the Medium. There are not, like the Light workload, applications added to increase or decrease the load.The actual resource usage of the workload has also changed (compared from Medium to Knowledge worker):
+
It also uses native Windows apps (Notepad and 7-Zip) to complete the print and zip actions used by the workload meta functions.
  
−
*There is a Disk Write increase of approximately 7%
+
==Power Worker==
−
*There is a Disk Read increase of approximately 3%
+
The Power Worker is designed for 2(v)CPU+ environments. This is a very intensive workload that puts maximum stress on the system, resulting in very high CPU, RAM and IO usage.
−
*There is an IOPS increase of approximately 13%
 
−
*There is a CPU usage increase of approximately 1%
 
  
−
==From Heavy to Power worker==
+
'''Important''': You must install the [https://www.loginvsi.com/documentation/index.php?title=Login_VSI_Pro_Library Pro Library] to be able to use this workload.
  
−
The Power worker is the new Heavy workload. In some scenario’s we discovered that the Heavy workload was not as heavy as we had designed. At some points (Storage) it was even less intensive as the Medium workload. This was not the desired result for a Heavy workload. The workload itself did not change in applications only in resources used. The changes are:
+
The Power Worker workload uses the following applications:
  
−
*Added copy and zip actions
+
*Adobe Reader
−
*Balanced the CPU usage
+
*Freemind/Java
−
*Increased the amount of Disk Reads
+
*Internet Explorer
−
*Increased the amount of Disk Writes
+
*MS Excel
 +
*MS Outlook
 +
*MS PowerPoint
 +
*MS Word
 +
*Photo Viewer
 +
*Simulated application install
  
−
The actual resource usage of the workload has been changed (compared from Heavy to Power worker)
+
For many of the applications in this workload, larger files and higher resolution media are used. It also uses native Windows apps (Notepad and 7-Zip) to complete the print and zip actions used by the workload meta functions.
  
−
*There is a Disk Write increase of approximately 30%
+
==Workload Resource Usage==
−
*There is a Disk Read increase of approximately 49%
+
We have created a table that contains the resource usage per workload, which provides crucial information for advanced users. These metrics are measured from a single user / VM, with all metrics based on the Knowledge Worker workload.
−
*There is an IOPS increase of approximately 54%
 
−
*There is a CPU usage decrease of approximately 4%
 
−
 
 
−
==Office worker==
 
−
There is a new workload added that has no precursor, called the Office worker. This workload is based on the Knowledge worker workload. The main goal of the Office worker workload is to be deployed in environments that use only 1vCPU in their VM’s. We have included the following applications into the Office worker workload:
 
−
 
 
−
*Outlook
 
−
*Microsoft Internet explorer
 
−
*Microsoft word
 
−
*Adobe reader
 
−
*Micrsoft Powerpoint
 
−
*Microsoft Excel
 
−
*Freemind (Java)
 
−
*Login VSI photo viewer
 
−
*Doro PDF Writer
 
−
 
 
−
Overall the Officeworker workload has less resource usage in comparison to the Knowledgeworker workload. The difference was calculated in the same environment with VM’s that had 2vCPU’s.
 
−
 
 
−
*There is no change in Disk Write
 
−
*There is Disk Read decrease of approximately 11%
 
−
*There is an IOPS decrease of approximately 13%
 
−
*There is a CPU usage decrease of approximately 22%
 
  
−
==Workload Resource Usage==
+
'''Note''': These values will be different depending on the hardware used, virtual machine optimizations and user concurrency. The information provided below is only provided to help you understand how they behave relative to one another.
−
We have created a table that contains the resource usage per workload. This will provide crucial information for advanced users. These metrics are measured from a single user / VM. All metrics are based on the Knowledge worker workload, this is our baseline in this table.
 
  
 
{| class="wikitable"
 
{| class="wikitable"
Line 105: Line 90:
 
'''IOPS'''
 
'''IOPS'''
 
|
 
|
−
'''Memory'''
+
'''RAM'''
−
|
 
−
'''vCPU'''
 
−
|-
 
−
|
 
−
Light
 
−
|
 
−
4.0
 
−
|
 
−
2
 
−
|
 
−
66%
 
−
|
 
−
52%
 
−
|
 
−
65%
 
−
|
 
−
5,2
 
−
|
 
−
1GB
 
−
|
 
−
1vCPU
 
−
|-
 
−
|
 
−
Medium
 
−
|
 
−
4.0
 
−
|
 
−
5-7
 
−
|
 
−
99%
 
−
|
 
−
93%
 
−
|
 
−
97%
 
−
|
 
−
7,4
 
−
|
 
−
1GB
 
−
|
 
−
2vCPU
 
−
|-
 
−
|
 
−
Heavy
 
−
|
 
−
4.0
 
−
|
 
−
8-10
 
−
|
 
−
124%
 
−
|
 
−
89%
 
−
|
 
−
94%
 
 
|
 
|
−
7
+
'''(v)CPU'''
−
|
 
−
1GB
 
−
|
 
−
2vCPU
 
 
|-
 
|-
 
|
 
|
 
Task Worker
 
Task Worker
 
|
 
|
−
4.1
+
4.1.x
 
|
 
|
 
2-7
 
2-7
Line 183: Line 111:
 
1GB
 
1GB
 
|
 
|
−
1vCPU
+
1(v)CPU
 
|-
 
|-
 
|
 
|
−
Office worker
+
Office Worker
 
|
 
|
−
4.1
+
4.1.x
 
|
 
|
 
5-8
 
5-8
Line 198: Line 126:
 
101%
 
101%
 
|
 
|
−
8,1
+
8.1
 
|
 
|
−
1,5GB
+
1.5GB
 
|
 
|
−
1vCPU
+
1(v)CPU
 
|-
 
|-
 
|
 
|
−
Knowledge
+
Knowledge Worker
 
|
 
|
−
4.1
+
4.1.x
 
|
 
|
 
5-9
 
5-9
Line 217: Line 145:
 
100%
 
100%
 
|
 
|
−
8,5
+
8.5
 
|
 
|
−
1,5GB
+
1.5GB
 
|
 
|
−
2vCPU
+
2(v)CPU
 
|-
 
|-
 
|
 
|
−
Power worker
+
Power Worker
 
|
 
|
−
4.1
+
4.1.x
 
|
 
|
 
8-12
 
8-12
Line 236: Line 164:
 
123%
 
123%
 
|
 
|
−
10,8
+
10.8
 
|
 
|
 
2GB
 
2GB
 
|
 
|
−
2vCPU+
+
2(v)CPU+
 
|}
 
|}
 +
 +
[[Category:Login VSI]]

Latest revision as of 11:40, 7 November 2017

This section describes the default workloads that come with the Login VSI product. These new and improved workloads feature a new timer mechanism, which is executed multiple times within each workload, and determines the performance of the system during testing.

Recommended: Simulating VDI Users - Introduction to Login VSI Workloads

Note: For more information on the additional workloads that are available to download, please see the Workload Add-ons section.

Task Worker

This is a light workload, which runs fewer application and starts/stops them less frequently than the other workloads, resulting in lower CPU, RAM and IO usage.

The Task Worker workload uses the following applications:

  • Adobe Reader
  • Internet Explorer
  • MS Excel
  • MS Outlook
  • MS Word

It also uses native Windows apps (Notepad and 7-Zip) to complete the print and zip actions used by the workload meta functions.

Office Worker

The main goal of this workload is to be deployed in environments that only use 1(v)CPU. It is based on the Knowledge Worker workload, but has less resource usage in comparison. However, it is more intensive than the Task Worker workload.

The Office Worker workload uses the following applications:

  • Adobe Reader
  • Internet Explorer
  • MS Excel
  • MS Outlook
  • MS PowerPoint
  • MS Word
  • Photo Viewer

It also uses native Windows apps (Notepad and 7-Zip) to complete the print and zip actions used by the workload meta functions.

Knowledge Worker

The Knowledge Worker is designed for 2(v)CPU environments. This is a well-balanced intensive workload that stresses the system smoothly, resulting in higher CPU, RAM and IO usage.

The Knowledge Worker workload uses the following applications:

  • Adobe Reader
  • Freemind/Java
  • Internet Explorer
  • MS Excel
  • MS Outlook
  • MS PowerPoint
  • MS Word
  • Photo Viewer

It also uses native Windows apps (Notepad and 7-Zip) to complete the print and zip actions used by the workload meta functions.

Power Worker

The Power Worker is designed for 2(v)CPU+ environments. This is a very intensive workload that puts maximum stress on the system, resulting in very high CPU, RAM and IO usage.

Important: You must install the Pro Library to be able to use this workload.

The Power Worker workload uses the following applications:

  • Adobe Reader
  • Freemind/Java
  • Internet Explorer
  • MS Excel
  • MS Outlook
  • MS PowerPoint
  • MS Word
  • Photo Viewer
  • Simulated application install

For many of the applications in this workload, larger files and higher resolution media are used. It also uses native Windows apps (Notepad and 7-Zip) to complete the print and zip actions used by the workload meta functions.

Workload Resource Usage

We have created a table that contains the resource usage per workload, which provides crucial information for advanced users. These metrics are measured from a single user / VM, with all metrics based on the Knowledge Worker workload.

Note: These values will be different depending on the hardware used, virtual machine optimizations and user concurrency. The information provided below is only provided to help you understand how they behave relative to one another.

Workload Name

VSI Version

Apps Open

CPU Usage

Disk Reads

Disk Writes

IOPS

RAM

(v)CPU

Task Worker

4.1.x

2-7

70%

79%

77%

6

1GB

1(v)CPU

Office Worker

4.1.x

5-8

82%

90%

101%

8.1

1.5GB

1(v)CPU

Knowledge Worker

4.1.x

5-9

100%

100%

100%

8.5

1.5GB

2(v)CPU

Power Worker

4.1.x

8-12

119%

133%

123%

10.8

2GB

2(v)CPU+

Retrieved from "https://loginvsi.com/documentation/index.php?title=Login_VSI_Workloads&oldid=5915"