Grouping of Computers

In PerformanceGuard there are several ways in which you can group computers that have PerformanceGuard agents installed. When computers are grouped, it becomes easy to compare performance data between locations or other dimensions, for example departments or users of particular applications.
One way of grouping happens automatically (but you can configure how it should happen). The automatic grouping means that you don't need to create any groups yourself in order for PerformanceGuard to work—you can create your own groups if and when you need them.
The other ways to group computers are manual. This lets you tailor your computer group structures to match the exact needs of your organization.
Let's take a look at the different ways to group computers in PerformanceGuard:
Automatic Grouping Based on Networks
PerformanceGuard automatically groups computers that belong to the same IP subnet into the same group. PerformanceGuard does this based on each computer's IP address and subnet mask.
In PerformanceGuard such a group is called a network group, because only computers that belong on a certain part of a network can be members of this group. Only PerformanceGuard can place computers in network groups.
Network groups are dynamic: If a computer is moved to another subnet (that is the computer's IP address information changes), PerformanceGuard will automatically move the computer to the network group that matches. If there isn't already a network group that matches, PerformanceGuard will automatically create one.
Can I control how PerformanceGuard places computers in network groups?  Yes, you can configure rules that'll determine how PerformanceGuard automatically places computers in network groups. With such rules you can split large networks into smaller ones, or you can join several small networks into a larger one (for example if you have a lot of VPN connections that you want to treat as a single network in PerformanceGuard). See Network Grouping.
Will I get a lot of empty network groups over time?  No. Consider network groups as an automated, dynamic environment: PerformanceGuard automatically adds network groups, but it also automatically removes network groups that are not needed anymore. You'll, however, still be able to view information about individual computers' historical memberships of such groups.
How do network groups work in a Citrix environment?  PerformanceGuard creates a special type of network group for Citrix servers. In Citrix environments, PerformanceGuard agents are installed on the Citrix servers rather than on individual computers, and PerformanceGuard will automatically create a network group for each of your Citrix servers. If your PerformanceGuard solution covers a mixed environment, you can have regular subnet-based network groups and Citrix server-based network groups side-by-side in the PerformanceGuard web interface.
To sum up the characteristics of network groups:

1. Automatic
2. Dynamic
3. Only PerformanceGuard can create them
4. Configurable rules if you want to control how PerformanceGuard automatically places computers in network groups
5. Special group variant for Citrix

Create Your Own Groups
You can only do this if you're a PerformanceGuard administrator.
Because PerformanceGuard automatically groups computers based on their network details, you don't need to set up groups yourself. However, it can sometimes be a good idea to create your own groups as a supplement to the automatically created network-based groups.
Example: Your organization consists of several regional offices. At each of the offices there are a few users who use a special finance application on their computers. You would like to be able to view performance data for users of this finance application. You create a group that contains the computers of those users. Now you can view performance data for that group, even though the computers are not on the same network.
You are able to do this because computers can be members of more than one group in PerformanceGuard. In other words, you can group your organization's computers by as many dimensions as you like. For example, you can group computers by user profiles. This can be useful because secretaries, accountants, salespeople and managing directors often use their computers differently. They may also have very different performance requirements, and you may want to measure their quality of experience separately.
In PerformanceGuard such a group that you create yourself is called a user-defined group.
User-defined groups are static: Computers will remain in user-defined groups until you remove them, even if computers are moved to another network. User-defined groups are completely independent of the automatically created network groups. This also means that—unlike the network groups—user-defined groups are not automatically removed if they become empty.
How do user-defined groups work in a Citrix environment?  When you create a user-defined group, PerformanceGuard asks you whether you want to create a regular user-defined group or a special user-defined group for Citrix servers (in Citrix environments, PerformanceGuard agents are installed on the Citrix servers rather than on individual computers). The difference is that only the special groups for Citrix servers are visible in the Citrix section of the PerformanceGuard web interface (ANALYZE > Citrix), whereas regular user-defined groups won't appear in the Citrix section.
Can I mix Citrix servers and regular computers in a group?  Yes, if it makes sense in your organization, you can include Citrix servers in a regular user-defined group, or include regular computers in a special user-defined group for Citrix servers.
To create your own groups in the PerformanceGuard web interface, select ADMINISTRATION > Computer Grouping > Groups, and then select the New Group tab.
To sum up the characteristics of user-defined groups:

1. Created manually
2. Static
3. No limit to the number of groups you can create
4. Computers can be members of more than one group
5. Special group variant for Citrix

Create Your Own Groups of Groups: Super Groups and Locations
You can only do this if you're a PerformanceGuard administrator.
You can group your groups (both network groups and user-defined groups) into larger logical units. In PerformanceGuard such larger units are called super groups or locations, depending on how you use them:

1. Super groups are simply logical containers that help you organize your other groups

You have a very high degree of flexibility, because:

1. You can include super groups and locations in other super groups or locations.
2. You don't need to have a single top node in your group hierarchy. You can easily have multiple top level groups (for example Europe and North America) rather than a single top level group (for example World) that would only serve to join the other groups.

Example: Click thumbnail to view image in full size.

To create your own locations or super groups, select ADMINISTRATION > Computer Grouping > Locations / Super Groups and then use the Create new group feature below the table of any existing groups.
To sum up the characteristics of super groups and locations:

1. Created manually
2. Static
3. No limit to the number of super groups or locations you can create
4. Can contain network groups, user-defined groups as well as other super groups and locations

A location must relate to at least one network group (i.e. it must contain a network group, or it must contain other locations or super groups that ultimately contain at least one network group).
Frequently Asked Questions About Computer Grouping
How are the groups described here different from agent configuration groups?  An agent configuration group is another type of group. Simply put, the groups described in the previous cover computers (that happen to have PerformanceGuard agents installed), whereas agent configuration groups cover the actual agents that are installed on the computers. Agent configuration groups control the configurations of the agents, that is which types of performance data that agents should collect, how often they should collect the data, etc. Each agent configuration group covers a number of agents, but an individual agent can only be a member of one agent configuration group so that it's always associated with exactly one configuration. This is different from the groups that contain computers, because a computer can be a member of multiple groups. A computer can be a member of multiple groups, but the agent installed on the computer can only be a member of one agent configuration group
I already have a group structure in Active Directory. Can I reuse that in PerformanceGuard?  You can't use Active Directory (AD) groups in place of network groups, user-defined groups, super groups or locations in PerformanceGuard. However, you can use your AD groups together with PerformanceGuard for two other purposes: 1) You can use AD group memberships to log in to the PerformanceGuard web interface with your domain users' AD credentials; see Active Directory. 2) You can automatically synchronize computers in AD groups to PerformanceGuard agent configuration groups; see Synchronize Computers in Active Directory Groups to Agent Configuration Groups.
Locations
For an overview of PerformanceGuard computer grouping concepts, see Grouping of Computers.
Select ADMINISTRATION > Computer Grouping > Locations to search for and manage groups of computers.
You can only do this if you're a PerformanceGuard administrator.
With Locations and Manual groups you can group other groups of computers into larger logical units, and view these larger logical units on the dashboard, in graphs, reports, etc.
What is a Location?
Locations can contain other sub-locations as well as Manual groups.
You can group your groups (both network groups and user-defined groups) into larger logical units. In PerformanceGuard such larger units are called Manual groups or locations, depending on how you use them. This gives you a very large amount of flexibility when you structure your group hierarchies.
For more information on how to create Locations, see Set Up Locations.
To sum up the characteristics of locations:

1. Created manually
2. Static
3. No limit to the number of locations you can create
4. Can contain Manual groups, Subnet groups and Citrix groups as well as locations

A location must relate to at least one network group (i.e. it must contain a network group, or it must contain other locations or Manual groups that ultimately contain at least one network group).
What is a User Defined Group?
If you have installed PerformanceGuard agents on your organization's computers, and the agents have begun reporting to PerformanceGuard, PerformanceGuard automatically sets up one or more computer groups for you. You can verify that by using the group search feature (see the following). The groups are based on the subnets that computers with agents belong to (see Network Grouping). This makes it easy for you to view data from your computers sorted by network/location.
Thus, you don't have to set up groups yourself—unless you find that setting up your own groups will bring you some benefits. Read the questions and answers in the following to find out if it makes sense to manually create some additional groups in your organization:
Why Have Groups of Computers?
Why is grouping of computers important?  Grouping is important because PerformanceGuard only keeps data for single computers for a short period of time. This is for performance and storage reasons. Computer data is aggregated to a group level, and old computer-specific data is deleted. You then only keep data at group level, so the more groups you create, the more data you get. By default all computers belong to two types of groups: the All Agents group and groups based on computers' IP addresses and subnets. PerformanceGuard uses the IP address and subnet information to automatically create network-based groups. This means that even with no manually created groups, your computers will still be grouped, because computers are automatically placed in the correct network group based on which IP subnet they currently belong to.
Why are network-based groups useful?  An example: PerformanceGuard measures response time on the computer so that you know what the actual user experience is, even with users at different physical locations with varying network bandwidth and latency. The response time is therefore a sum of network transport time from client computer to server, server response time, and network transport time for the first byte of the response to arrive back at the client. Computers at a location will typically be on the same subnet, and they will thus belong to a specific network-based group. This means that you can view response times for different physical locations simply by viewing the measurements aggregated on the network-based groups. One day users at the London location complain that the system is slow. You haven't had any complaints from users at the Manchester location. You compare the response times from the network-based groups that represent the two locations. This way you can find out if both locations experience long response times or if it's only a problem at the London location, and then you'll know if there is a network/computer problem or a backend problem.
Traveling Users
What if users often move between networks?  Users on the move, for example salespeople, will often use many different networks, and their computers will thus switch between different network-based groups in PerformanceGuard. You can keep track of individual computers' historical memberships of such groups.
Create Your Own Groups
You can create groups manually as well, for more information on how to create groups manually, see Set Up Locations.
Why would I want to create groups manually?  It can sometimes be a good idea, for example if you have a user group that isn't on the same network. It could be a small number of people at each of your locations that use a special application. In that case you can create a group containing just those users' computers. In other words, you can group by user profiles, and this can be useful because accountants, salespeople and managing directors may use their computers differently.
If I create my own groups, can a computer belong to more than one group?  Yes, a computer can be member of multiple groups. This lets you group your computers by as many dimensions as you like.
Can I create groups of groups?  Yes, you may create groups of groups. In PerformanceGuard these are called super groups. You may, for example, create super groups of network-based groups in order to group all locations in a country or state. Because super groups may contain groups as well as other super groups, you can then create groups for entire regions or continents.
Search for Groups
Use some or all of the following parameters in your search:

1. Group name: The search is case insensitive, so entering on would return the groups Lisbon as well as Ontario.
2. Group type: Only show the selected group type.
3. Minimum number of agents: Minimum number of computers in the group.
4. Online agents: Show a row with the number of computers online in the particular group.
5. Citrix Sessions: The number of Citrix sessions with activity during the last report delivery interval.
6. Application Pings: Show the number of application pings made by computers in the group during the last report delivery interval.
7. Trace routes: Show the number of traceroutes made by computers in the group during the last report delivery interval

Search Result

1. Show a list of all members of the group.
2. Edit the group.
3. Business Unit: The name of the business unit that the group belongs to.
4. Name: The name of the group.

You can change the name. Example: You think that the automatically generated name Net 192.168.101.155/10.0.30.0/24—which, by the way, means computers on the 10.0.30.0/24 network behind router 192.168.101.155—isn't meaningful enough, so you click the edit icon and change the group's name to Manchester production network.

1. Description: A supplementary description of the group.
2. Network: The group's network address.
3. Router: The router address, if relevant.
4. Agents: Number of computers in the group.

Create New Group
To create a new group:

1. Select the New Group tab.
2. Select whether you want to create a regular group (Agent group) or a group for Citrix servers (Citrix server group).
3. Select required business unit if business units are used in your organization.
4. Specify a Name (required) and a Description (optional).
5. Click the Create group button.
6. When ready, click the icon to search for and add computers to your newly created group.

Network Grouping
For an overview of PerformanceGuard computer grouping concepts, see Grouping of Computers.
PerformanceGuard automatically places computers that belong to the same IP subnet in the same group. This type of group is called a network group because only computers that belong to a certain IP network can be a member of this group.
Only PerformanceGuard can place computers in these groups, based on the IP addresses and subnet masks of computers that have PerformanceGuard agents installed. If a computer is moved to another subnet (that is the computer's IP address changes), PerformanceGuard will automatically move the computer to the corresponding network group.
PerformanceGuard automatically removes network groups that are not needed anymore (that is if there are no computers that belong in the network groups anymore). You'll, however, still be able to view information about individual computers' historical memberships of such groups.
Automatic Network Groups
If a suitable network group doesn't exist when the PerformanceGuard agent on a computer reports to PerformanceGuard, PerformanceGuard automatically creates a new network group that fits the IP address and netmask of the computer.
The new group is by default called Net a.b.c.d/x where a.b.c.d is the network address and x is the netmask length of the network.
Set Up Network Grouping Rules
You can only do this if you're a PerformanceGuard administrator.
There may be situations where you want to control the creation of new network groups. You can do this by setting up network grouping rules: Select ADMINISTRATION > Computer Grouping > Network Grouping and then select the Create New tab.
A rule consists of a network, a minimum netmask length and a maximum netmask length.
The network identifies the IP addresses for which the rule applies. You specify the network as a network address and a netmask length.
The minimum netmask length field controls the maximum size for network groups within the specified network by setting a lower limit for the network group mask length. The maximum netmask length controls the minimum size by setting an upper limit for network group mask length.
The number of network groups resulting from a single network grouping rule can at most be 2^(maximum netmask length - network netmask length).
Example: You specify a network grouping rule with a network of 172.5.0.0/22, a minimum netmask length of 22 and a maximum netmask length of 24. You then get the following network groups:172.5.0.0/24 (for computers with IP addresses in the range 172.5.0.0-172.5.0.255)172.5.1.0/24 (for computers with IP addresses in the range 172.5.1.0-172.5.1.255)172.5.2.0/24 (for computers with IP addresses in the range 172.5.2.0-172.5.2.255)172.5.3.0/24 (for computers with IP addresses in the range 172.5.3.0-172.5.3.255)
Join Small Networks into Larger One (Useful for VPN Connections)
Imagine that you specify a network grouping rule with the network 10.2.4.0/22 and a maximum netmask length of 24. That would mean that within the network 10.2.4.0/22 no network group can be created that has a netmask that's higher than 24.
Now, if the PerformanceGuard agent reports an IP address of 10.2.5.130 and a netmask length of 26, you would expect that the computer would be placed in a network group called 10.2.5.128/26.
However, because of your rule the network group will be 10.2.4.0/24. This is because the computer's IP address falls within the network 10.2.4.0/22, and there's an upper limit of 24 on the netmask length.
This kind of rule is useful if you have a lot of small networks (for example VPN connections) that you want to see as one larger network.
Split Large Networks into Smaller Ones (Useful for Networks Divided into Smaller VLANs)
Imagine that you specify a network grouping rule with the network 10.2.0.0/16 and a minimum netmask length of 24. That would mean that within the network 10.2.0.0/16 no network group can be created that has a netmask that's lower than 24.
Now, if the PerformanceGuard agent reports an IP address of 10.2.5.130 and a netmask length of 16, you would expect that the computer would be placed in a network group called 10.2.0.0/16.
However, because of your rule the network group will be 10.2.5.0/24. This is because the computer's IP address falls within the network 10.2.0.0/16, and there's a lower limit of 24 on the netmask length.
This kind of rule is useful if you have one big network divided into smaller VLANs.
Explicitly Define Your Network Groups (Useful for Citrix ICA Clients)
If you specify a rule with the network 10.2.4.0/22 and a maximum netmask length of 22, it means that within the network 10.2.4.0/22 one network group can be created.
So, if you specify a maximum netmask length that's equal to the netmask length of the network, you have explicitly defined a network group.
This kind of rule is useful if you only have Citrix ICA Clients in the network. ICA Clients only report their IP addresses and not their netmasks, so PerformanceGuard won't really know which network groups the ICA Clients should be placed in. Therefore, PerformanceGuard by default assigns a netmask length of 24 to ICA Clients if they don't fit into any existing network group. By explicitly defining your network groups, you can make sure that ICA Clients will be grouped correctly.
Computers Behind NAT-Enabled Routers
Computers behind NAT (Network Address Translation)-enabled routers usually have IP addresses in private address ranges, such as 192.168.0.0/16 and 10.0.0.0/8. Because of this you may experience that computers on different physical locations have the same private IP addresses configured. To avoid conflicts, PerformanceGuard will group the computers based on the public IP address of the router instead.
The public IP address of the router is defined as the client endpoint of the TCP communication between the server that runs PerformanceGuard and the PerformanceGuard agent. If you have several NAT-enabled routers behind each other, PerformanceGuard will only recognize the first public router.
Whether a router is NAT-enabled or not is decided by comparing the client endpoint IP address and the reported agent IP address. If they don't match, PerformanceGuard assumes that Network Address Translation has taken place.
In PerformanceGuard such network groups will be called something like Net 123.76.76.42/192.168.101.0/24, which means computers located on private subnet 192.168.101.0/24 behind a router with a public IP address of 123.76.76.42.
Unwanted Network Groups
Automatically generated network groups can't be deleted from within the PerformanceGuard web interface. This is because they would be recreated by PerformanceGuard if you don't set up appropriate network grouping rules to prevent the creation of the unwanted network groups. Thus, the solution is to set up network grouping rules that'll prevent the automatic creation of the network groups that you don't want.
PerformanceGuard automatically removes network groups that are not needed anymore (that is if there are no computers that belong in the network groups anymore). You'll, however, still be able to view information about individual computers' historical memberships of such groups.