Scalability and performance
Understanding the scale of your XenMobile infrastructure plays a significant role in how you decide to deploy and configure XenMobile. This article contains data from scalability tests and guidance on determining infrastructure requirements for performance and scalability for small- to large-scale, on-premises XenMobile enterprise deployments.
Scalability is defined here in terms of the ability of devices already enrolled in the deployment to reconnect to the deployment at the same time.
Scalability is defined as the maximum number of devices enrolled in the deployment.
Login Rate is defined maximum rate at which existing devices can reconnect to the deployment.
The data in this article are derived from testing on deployments ranging in size from 10,000 to 75,000 devices. The tests comprised mobile device using known workloads.
All testing was done on XenMobile Enterprise edition.
Testing was done using the NetScaler Gateway 8200. NetScaler appliance with similar or greater capacity can be expected to produce similar or greater scalability and performance.
A summary of scalability test results follows.
Summary of Scalability test results for deployments of up to 75,000 devices
Login rate (reconnection rate of existing users) — Up to 9,375 devices per hour
XenMobile Enterprise Edition
XenMobile Server 7-node cluster
Database: Microsoft SQL Server external database
Test results by device population and hardware configuration
|Number of devices||12,500||30,000||60,000||75,000|
|Reconnection rate of existing devices per hour||1,250||3,750||7,500||9,375|
|XenMobile Server – mode||Standalone||Cluster||Cluster||Cluster|
|XenMobile Server – cluster||N/A||3||5||7|
|XenMobile Server – virtual appliance||Memory = 8 GB RAM; vCPUs = 4||Memory = 16 GB RAM; vCPUs = 6||Memory = 24 GB RAM; vCPUs = 8||Memory = 24 GB RAM; vCPUs = 8|
|Active Directory||Memory = 4 GB RAM; vCPUs = 2||Memory = 8 GB RAM; vCPUs = 4||Memory = 16 GB RAM; vCPUs = 4||Memory = 16 GB RAM; vCPUs = 4|
|Microsoft SQL Server external database||Memory = 8 GB RAM; vCPUs = 4||Memory = 16 GB RAM; vCPUs = 8||Memory = 24 GB RAM; vCPUs = 16||Memory = 24 GB RAM; vCPUs = 16|
|Active Directory Configuration||Profile used|
|Levels of nesting||5|
|XenMobile Server Configuration||Total||Per user|
|Web and SaaS||20||20|
|Active Directory groups per delivery group||10|
|Number of databases||1|
Device connections and app activities
These scalability tests collected data on the ability of devices enrolled in a deployment to reconnect over an 8-hour period.
The tests simulated a reconnect interval during which reconnecting devices obtain all entitled security policies, subjecting XenMobile Server nodes to higher than normal load conditions. During subsequent reconnections, only changed or new policies are pushed to iOS devices, lessening the load on the XenMobile Server nodes.
These tests used a mix of 50 percent iOS devices and 50 percent Android devices.
These tests assume that the reconnecting Android devices have received prior GCM notifications.
During the 8-hour test interval, the following app-related activities occurred:
Secure Hub was opened once to enumerate entitled apps
2 SAML web apps were opened
4 MAM apps were downloaded
1 STA was generated for use by Secure Mail
240 STA ticket validations, one for each Secure Mail reconnect event over a micro VPN, were performed.
For the reference architecture for deployments used in these scalability tests, see “Core MAM+MDM Reference Architecture” in Reference Architecture for On-Premises Deployments.
Caveats and limitations
Note the following when considering the scalability test results in this article:
Windows platform was not tested.
Policy push was tested for iOS and Android devices.
Each XenMobile Server node supports a maximum of 12,000 devices simultaneously.