The SSID Access Control tab contains settings that control access to the SSID, for example, Firewall and Client Authentication settings.
You can configure the following firewalls on the Access Control tab:
Note: You can not enable firewall settings if Dynamic VLANs is enabled under CONFIGURE > SSID >Security > 802.1X.
To configure the firewall settings, see Configure Firewall Settings.
You can enable Apple's
Bonjour Gateway feature that allows access to Apple devices on the network.
Note: Bonjour Gateway does not work when the Network is set to NAT mode. If you have set the Network to NAT mode, CV-CUE grays out Bonjour Gateway and prompts you to change the Network setting from within the Access Control tab.
For details, see
How Arista Supports Bonjour Gateway. To configure Bonjour Gateway, see
Configure Bonjour Gateway.
You can enable
Redirection to redirect either Smartphones & Tablets or all clients of the SSID to the
Redirect URL that you specify. This could be useful, for example, in an enterprise network where you might want smartphones and tablets to be redirected when accessing the SSID, but allow laptops and desktops to directly start using Wi-Fi. You can also have a
Walled Garden of sites that the user can access before login.
Note: You must enter at least the Redirect URL in the Walled Garden field, since the user must be able to access that URL before login.
To configure Redirection, see Configure Redirection in SSID Access Control.
Organizations such as enterprises and educational institutions (K-12 and higher education) often implement a centralized Authentication, Authorization and Accounting (AAA) management to enforce Role Based Control , also called Role Based Access Control (RBAC). RBAC enables network administrators to restrict system access to authorized users. Users are granted controlled access to network resources based on the roles assigned to them or the groups to which they belong. Typically, organizations implement this kind of controlled access by using RADIUS. When users connect to the network, they are first authenticated and then authorized to access appropriate resources on the network.
In the case of a WLAN network, user access restrictions could mean that only specific VLANs or a fixed bandwidth is provided to users based on the user roles defined in the RADIUS server. You can also enforce which applications a user can access over the WLAN network based on the user role.
Arista uses Role Profiles to define various WLAN access roles, and to create RADIUS Vendor Specific Attribute (VSA) based rules and Google Organizational Unit (OU) rules to authorize Wi-Fi users. A network administrator can define various role profiles that specify the restrictions to be placed on the Wi-Fi user to whom the profile is assigned. The administrator can then define multiple VSA rules (for RADIUS) or Google OU rules (for Google Integration) here in SSID Access Control, and assign role profiles through these rules to the Wi-Fi users that connect to the SSID.
Let us consider an example. When you define a Rule Type for RBAC, then the OU returned from Google or the role obtained from the RADIUS VSA must contain the string entered in the Enter Value field. For example, if the string in the Enter Value field is ‘/*/Elementary School/*/Student’, then this will match with ‘/SJUSD/Elementary School/Almaden Elementary/Student’ in Google/VSA.
It could happen that you have different settings in the SSID tabs and different ones in the Role Profiles tab. What happens then? For the answer, see Role Profile.
To configure Role Based Control, see Configure Role Based Control.
To control clients that can access this SSID, you can create Allow and Deny lists of client MAC addresses. See How the Client MAC Allow and Deny Lists Work and Requirements for details on the feature.
With Client Isolation enabled on an SSID, Wi-Fi clients associated with the SSID are allowed to communicate only with their gateway; they cannot communicate directly with any other hosts on the same subnet—including other clients on the same SSID, clients associated with other SSIDs on the same subnet, and hosts connected to the wired network on the same subnet. An AP running an SSID with Client Isolation discards all packets from a client if the destination IP address is on the same network as the client, except for packets destined to the gateway.
Consider two Wi-Fi clients, Client 1 and Client 2, associated with different APs on the same SSID, SSID 1. As shown in the figure below, with Client Isolation enabled, AP1 discards packets originating from Client 1 and destined to Client 2.

If NAT is enabled on an SSID, an AP running the SSID becomes the gateway node of its own private subnet. Consider Client 1 and Client 2 in the figure above. If these clients are associated with a NAT-ed SSID, they cannot see each other’s IP address. Thus, it is NAT rather than Client Isolation that prevents direct connections between these clients; Client Isolation prevents direct connections only between clients of the same AP.
Note that even with a NAT-ed SSID, the net effect of enabling Client Isolation is the same as in the case of a bridged or tunneled SSID: clients on the same SSID cannot communicate directly with each other. But the mechanisms that prevent such communication are different: NAT prevents direct communication between clients on different APs and Client Isolation prevents direct communication between clients of the same AP.
Client Authentication adds another layer of security to your network. It authenticates
clients, i.e. user devices, in addition to mechanisms configured in the SSID
Security tab that authenticate
users (e.g. WPA2-PSK). Client Authentication uses either
Google Integration or
RADIUS MAC Authentication. See
Google Integration for more information.
Note: If you have configured 802.1X authentication in the SSID Security tab, then CV-CUE grays out the RADIUS MAC Authentication option, since 802.1X already is a RADIIUS-based mechanism.
Some Wi-Fi clients send Diassociation messages whenever they enter a "sleep" mode. If the AP immediately sends an Accounting Stop request to the RADIUS server, the RADIUS server clears the client info and the client has to reauthenticate when it wakes up. This could cause frequent and unnecessary reauthentication. The
Accounting Stop Delay is the number of minutes that the AP waits between the time it receives the Disassociation and the time it sends the Accounting Sopt message to the RADIUS server. If the client wakes up in the interim and communicates with the AP, the Accounting Stop message is not sent and the client does not need to reauthenticate.
For the other RADIUS settings, see Configure SSID Security Settings.
You can choose to either Disconnect or Stay Connected and Assign Role to the user. To assign a role, you need to select one from those defined on the Role Profile tab. You might configure Client Authentication before you have created any Role Profile. When you click Add / Edit under Select Role, a window appears in the right pane, allowing you to define a Role Profile without having to leave Client Authentication.
To configure Client Authentication, see Configure Client Authentication.
L3-4 Firewall
Arista Access Points (APs) have firewall capabilities. The AP firewall monitors the traffic passing through the AP and takes actions based on user-defined rules.
The firewall is stateful, that is to say, it keeps track of whether the connection has been opened in the outgoing direction (wireless to wired-side) or in the incoming direction (wired-side to wireless), and takes appropriate actions on the packets based on the direction in which the connection was opened. The following image illustrates the conventions used for directions.
Note that this is not the Internet facing firewall. Its main purpose is to facilitate traffic controls, such as allowing/disallowing access to certain assets and/or applications for wireless users. The firewall rules are defined and enforced on a per SSID basis. Arista APs support multiple SSID profiles, thereby enabling multiple firewall configurations to co-exist.
The following use cases illustrate typical applications for the Arista AP firewall functionality:
- Block guest Wi-Fi users from accessing the private/corporate subnet. This serves as an additional security control to ensure that guest Wi-Fi users can access only public Internet and nothing in the private address space.
- Block or allow access to specific domain names.
- Allow guest Wi-Fi users to access only HTTP and HTTPS content in the Internet. This is typically done to control the type of traffic guest users can generate.
- Implement DNS-based content filtering to prevent access to non-family-friendly web sites, security threats, and peer-to-peer file sharing. The firewall can be used to ensure that Wi-Fi clients necessarily use the specified content filtering DNS server, such as Norton ConnectSafe, and cannot bypass it.
- Enforce use of IPsec VPN for wireless clients.
Note:
- When you enable L3-4 Firewall Rules, you can see the default rule Action : Block on the UI. If you enable L3-4 Firewall Rules and do not define any rules at all, the default rule applies, i.e., all traffic is blocked.
- The AP compares traffic with rules from top to bottom until it finds the first match. Once it finds the first match, the AP does not compare the rest of the rules. If it finds no match with any of the defined rules, the AP uses the default rule at the end. You can re-order the rules using the drag-and-drop feature to reposition them at the desired level.
In case of a conflict between rules on the L3-4 Firewall and those on the Application Firewall, the AP decides using this Decision Table.
Example Use Case of L3-4 Firewall
Let us look at a rule set that might be found on a Guest SSID in a retail store deployment.
Goal for Retail Store: Allow only HTTP/HTTPS Internet access, with content filtering and no access to private subnets.
Table 1. Example Rules Table for Retail Store
Rule Number |
Rule Name |
IP / Hostname |
Port |
Action |
Protocol |
Direction |
1 |
Content Filtering DNS1 |
199.85.126.30 |
53 |
Allow |
UDP |
Outgoing |
2 |
Content Filtering DNS2 |
199.85.127.30 |
53 |
Allow |
UDP |
Outgoing |
3 |
Block All Other DNS |
* |
53 |
Block |
UDP |
Outgoing |
4 |
No Local Access |
192.168.0.0/16, 172.17.0.0/21, 10.0.0.0/8 |
|
Block |
Any |
Any |
5 |
Allow HTTP / HTTPS |
* |
80, 443 |
Allow |
TCP |
Outgoing |
6 |
Default |
|
|
Block |
|
|
Rule 1 - Allow outbound UDP port 53 to Content Filtering (Norton) DNS1/199.85.126.30. This rule implements DNS-based content filtering to block access to web sites that contain non-family-friendly content, pose security risks, and promote file sharing applications. DNS uses UDP port 53. So this rule allows outgoing UDP connections destined to port 53 on a content filtering DNS server with the 199.85.126.30 host IP address.
Because the firewall is stateful, the return path is automatically allowed and you do not need a separate rule for the return path. This is true for the other rules as well.
Rule 2 - Allow outbound UDP port 53 to Content Filtering (Norton) DNS2/199.85.127.30. Like Rule 1, this rule also implements DNS-based content filtering. This rule provides DNS server redundancy.
Rule 3 - Block all outbound UDP 53. This rule blocks all DNS traffic excluding that which is allowed by Rules 1 and 2. This rule prevents users from statically configuring DNS server addresses on their clients to circumvent content filtering.
Rule 4 - Block traffic to destination 192.168.0.0/16, 172.17.0.0/21 and 10.0.0.0/8. Blocks access to private/corporate subnets. This rule blocks any wireless traffic addressed to any host in the 192.168.0.0/16, 172.17.0.0/21 and 10.0.0.0/8 subnets. The Protocol specified for this rule is Any, which covers any protocol carried over IP. Because there are protocols that do not implement the port concept (e.g. ICMP), the port number gets grayed out when Any is selected as protocol. This rule is ideal for restricting users on the Guest Wi-Fi from accessing private subnets.
Rule 5 - Allow any traffic outbound to TCP port 80, 443. Allow clients to open outgoing TCP connections to port 80 (allows outgoing HTTP connections) and allow clients to open outgoing TCP connections to port 443 (allows outgoing HTTPS connections). The wildcard character (*) represents “any” hosts.
Rule 6 - Default rule is set to Block, which means that all other kinds of communication, except the ones enabled by the rules 1-5, are disallowed.