Recommended Configuration
App Portal / App Broker 2024 R1
In this section, the basic recommended configuration for an App Portal installation is provided, as well as the steps you can take to perform an advanced configuration analysis.
Basic Recommended Configuration
The following table provides a summary of the recommended App Portal / App Broker configuration, including hardware sizing recommendations.
| Component | Development / Lab Environment | Production Environment |
|---|---|---|
| Server Count | 1 | 1 |
| Server Role | App Portal / App Broker Web Server | App Portal / App Broker Web Server |
| Processor | 1 processor, 2 core | 2 processors, 4 core |
| Memory (RAM) | 8 GB | 32 GB |
| Hard Disk Space | 70 GB | 70 GB |
| Database sizing recommendations | 50 GB (based on estimated growth size of the deployment technology database of 500 GB) |
The App Portal / App Broker web service can only be enabled and running on a single web server at a time.
It is recommended that the App Portal / App Broker database reside on its own Microsoft SQL Server. It is not necessary to install App Portal / App Broker on the same SQL Server as the System Center Configuration Manager or Altiris.
Advanced Configuration Analysis
To determine the recommended system resource requirements for an App Portal installation, you should first gather metrics to calculate the base level system resource requirements, and then compare those metrics against data gathered from production implementations of App Portal for a company that is similar in size.
Initial Platform Sizing
Based on the number of devices that App Portal will interact with at your enterprise, determine your company’s system resource tier.
| Tier | # of Devices | Processor | Memory (RAM) | Storage (Hard Disk) |
|---|---|---|---|---|
| T1 | 1 – 9,999 | 1 Processor, 4 Core | 8 GB | 70 GB |
| T2 | 10,000 – 49,999 | 1 Processor, 8 Core | 16 GB | 80 GB |
| T3 | 50,000 – 99,999 | 2 Processor, 4 Core | 32 GB | 90 GB |
| T4 | 100,000 – 399,999 | 2 Processor, 8 Core | 64 GB | 100 GB |
| T5 | 400,000+ | Special |
Common Performance Bottlenecks
Common bottlenecks for web server performance include network bandwidth, CPU, memory, and I/O. Another important consideration for web servers is whether the content type is static or dynamic. The following table lists common bottleneck areas. Review this list to determine if any of these bottlenecks apply to your organization.
Common Bottleneck Areas
| Bottleneck Area | Characteristics Reviewed |
|---|---|
| Network Sizing | The amount/frequency of a request sent to the server (hits) The size of that request (client request) The size of the request (web server response), total size of average web page, including all objects. |
| CPU Sizing | Steady state of less than 80% CPU utilization assured Busy time and observation time Transactions during observation time |
| Memory Sizing | OS and web server memory usage Generating dynamic content Calculating dynamic and static content ratios OS and web server caching |
| I/O Sizing | Amount of disk space, not to exceed 85% of used and 15% free space Amount of data |
Comparative Analysis
In this step, compare the system resource tier identified in Initial Platform Sizing with several customer production implementations of App Portal which are of similar size, and determine whether the your organization is similar to one of them.
| Customer ID | Brief Description | System Resources |
|---|---|---|
| Customer A | Multinational medical device, pharmaceutical, and consumer packaged goods company with approximately 250,000 users and 300,000 managed devices. Customer A has approximately 10,000 catalog items managed within the App Portal catalog, connected to FlexNet Manager Suite. | Tier 4 2 Processors, 8 Cores 64 GB RAM 200 GB Hard Disk Storage |
| Customer B | Multinational consumer goods company with approximately 150,000 users and 175,000 managed devices. Customer B has approximately 4,000 catalog items managed within the App Portal catalog, connected to FlexNet Manager Suite. | Tier 3 with Additional Disk Storage 2 Processors, 4 Cores 32 GB RAM 150 GB Hard Disk Storage |
| Customer C | Regional American bank holding company with approximately 65,000 users and 80,000 managed devices. Customer C has approximately 2,500 catalog items managed within the App Portal catalog. Customer C does not have FlexNet Manager Suite. | Tier 5 (Special) 2 Processors, 8 Cores 128 GB RAM 250 GB Hard Disk Storage |
Formulas Used
There are a number of formulas used to identify which system resource tier a solution should implement.
Network Calculation
At a high level, basic network calculation can be expressed as follows:
$b = h x s$
where:
- b = the required number of network bandwidth hits (bits per second)
- h = the number of web server hits per second
- s = the average size of each hit in bits
Network Transfer Rate
To calculate the network transfer rate, use the following calculation:
$\frac{({bit}{per}{second})}{8} = \frac{MB}{{MB}{per}{second}}$
Number of Hits Per Day
If you have access to web traffic information, you can estimate the number of hits per day. In this case, break the day into four-hour windows, and select the number of hits in the highest four-hour window. Multiply this figure by 6. Then divide the result by 24 to determine number of hits per hour.
$\frac{h x 6}{24} = {hits}{per}{hour}$
where:
- h = the number of hits in the highest four-hour window.
CPU Utilization
The formula to determine CPU utilization is as follows:
$U = \frac{b}{t}$
where:
- U = utilization
- b = busy time
- t = observation time
Lower Memory Limit
The formula to calculate the lower memory limit is as follows:
$M = o+{(r} x{d)}+{(r} x{g)}+ f+ w$
where:
- M = Minimal acceptable amount of memory/sec
- o = OS and web server memory requirements, taking into account concurrent connections/sec
- r = Number of requests for dynamic content/sec
- d = Memory cost to generate dynamic pages
- g = Memory cost to generate static pages
- f = OS file system caching, memory allocated
- w = Web Server caching, memory allocated
Disk Space
The formula to calculate disk space sizing is as follows:
$S = d+{(d} x{0.15)}$
where:
- S = disk space
- d = amount of data