--- title: "Tunnel architectures" slug: "tunnel-architectures-v7-0" description: "Explore AppGate ZTNA's tunnel architectures, enabling secure multi-site connections and advanced routing for optimal network performance and security." tags: ["Tunneled Traffic", "Tunnel Architectures", "Tunneling"] updated: 2026-09-21T20:36:59Z published: 2026-09-21T20:36:59Z canonical: "support.appgate.com/tunnel-architectures-v7-0" --- > ## Documentation Index > Fetch the complete documentation index at: https://support.appgate.com/llms.txt > Use this file to discover all available pages before exploring further. # Tunnel architectures ## Understanding tunnel architectures in the AppGate ZTNA Collective **Site** is an AppGate ZTNA term for the environment that the system protects. Do not confuse it with a physical site. A Site may represent one or more hosts in a subnet, a DMZ, a geographic location, or a business unit. Sites typically include one or more Gateways (see [High availability](/v1/docs/high-availability-ha-v7-0)) to manage the client tunnels for that Site. Unlike traditional VPN solutions, clients can connect to multiple Sites in a single Collective at the same time. New Collectives include a default Site preconfigured to use [entitlement-based routing](/v1/docs/routing-client-traffic-v7-0) (EBR). Edit this default Site to suit your environment. To start, you may want to add only a DNS resolver and leave the remaining settings at their defaults. Every entitlement you configure must relate to a default Site—that entitlement’s home. This relationship lets the multi-tunnel network driver establish a tunneled connection to each Site for which the user holds an entitlement token. When the client connects to a Gateway, it sets up routes to direct traffic down the appropriate tunnel. The Gateway then routes the tunneled traffic to the protected network. ### The tunnel AppGate ZTNA supports both TCP and UDP for the tunnel. This choice does not affect the protocols supported within the tunnel. Clients establish a secure tunnel to an available Gateway on each Site, based on preset weighting. The multi-tunnel network driver receives an IP address from the IP pool, so tunneled client-to-Gateway connections appear as standard network devices. The tunnel supports TCP, UDP, GRE, and ICMP, as well as inbound and outbound traffic. This support enables more complex deployments such as IP telephony. Tunnels use the default MTU of 1500 for the appliances' external NICs. On mobile operating systems, each tunneled connection requires one instance of the operating system VPN service per Site. Running multiple VPN services consumes significant resources, so minimize the number of Sites when deploying to mobile clients. > [!NOTE] > NOTE > > On iOS and iPadOS, do not use more than five Sites. Exceeding five Sites is likely to exceed the memory limit for VPN services. The default tunnel mode is TLSv1.3. QUIC is also available for use cases that require it. #### Using QUIC for the tunnel protocol Your choice of tunnel protocol does not affect which protocols the tunnel supports between the client and the resource. QUIC has no handshaking. It streams packets with no guarantee of packet delivery or retry on failure. This can improve performance on high-latency networks where low delay matters more than packet loss—for example, VoIP traffic. Consider the following when using QUIC: - To use single packet authorization (SPA) functionality on Gateways, use Check UDP (and TCP) SPA mode. In Check TCP SPA mode, SPA applies only to TLS-based connections such as the client-to-Controller connection. It has no effect on the QUIC tunnel to the Gateways. Tunneled connections follow the software-defined perimeter (SDP) network model. AppGate ZTNA also supports advanced connectivity options for more complex environments. ### Tunnel connection models Gateway and Site configurations include advanced connectivity options that control how tunneled connections are handled. Use these options to meet specific requirements, such as VPN replacement or optimal geographic routing. These options are useful when deploying AppGate ZTNA in environments not originally designed to the SDP model. Outside the SDP network model, use the connectivity options for advanced use cases only. Some models must be enabled in policies so that assignment criteria can limit which users receive non-standard connection handling, such as system administrators. #### Software Defined Perimeter network model The entitlement tokens that the client receives include the *Client Hostname/IP* of the Gateways (120.33.17.10,11,12,20) configured in [Secure Tunnel Settings](/v1/docs/adding-a-gateway-v7-0#secure-tunnel-settings). Most users connect to only the 111.x Site, but a couple connect to both the 111.x Site and the 112.x Site. By default, *Disable Source NAT on Gateways* is not enabled, so Gateways perform source NAT from the 192.168.100.64-254 address range to the 172.17.111.x and 172.17.112.20 IP addresses of the respective Gateways. Each user automatically receives a unique source port per TCP/UDP stream, providing compatibility with most network environments. ![](https://cdn.us.document360.io/e7482ce5-acfa-4c3b-b76c-b6c32472644a/Images/Documentation/sdp_network_model_appgate.png?sv=2026-02-06&spr=https&st=2026-11-21T21%3A21%3A35Z&se=2026-11-21T22%3A21%3A35Z&sr=b&sp=r&sig=lCqRElFlfKmqwuFcWH9PlE4vQ2xAoOk2WJ%2Fmet%2Fbkr0%3D) #### Using an external load-balancer External load balancers offer advanced high availability (HA) features similar to those built into the system, and you can use them in front of both the Controller and Gateway. However, two core security features—single packet authorization (SPA) and mutual TLS (mTLS)—significantly limit which load-balancer features are available. Some advanced HA features are not supported at all. When you use an external load balancer, you cannot enable UDP-TCP SPA mode because the load balancer cannot handle both UDP and TCP traffic. Load balancers also lack an equivalent cloaking feature, making them visible on the internet. This visibility defeats a core principle of the SDP model. Using TCP SPA mode with mTLS also limits your load-balancer options. These two security features protect against man-in-the-middle attacks, which makes application-layer load balancers unusable. The load balancer must operate exclusively at the DNS or TCP/IP layer. Round-robin DNS is one option. Edit the *Client Hostname/IP* (Gateway) or the *Profile DNS name* (Controller)—issuing different values to user groups in different locations, for example. The DNS service selects an IP address using round-robin or predefined criteria such as geolocation. Configure the load balancer to use the Healthcheck Server to probe each appliance, so that clients are not directed to unavailable appliances. In this scenario, the client has no list of IP addresses. If it receives a timeout from the Controller, the user cannot authenticate. A TCP/IP layer load balancer is the other option. Set the DNS record for *Client Hostname/IP* (Gateway) or the *Profile DNS name* (Controller) to point to the load balancer. The load balancer becomes the only hostname or IP address the client connects to. The load balancer forwards traffic to an appliance based on predefined criteria. Configure it to use the Healthcheck Server to probe each appliance, so that clients are not directed to unavailable appliances. In this scenario, the client has no list of IP addresses. If it receives a timeout from the Controller, the user cannot authenticate. If you configure the load balancer with Controllers to direct all traffic to the primary Controller first, significantly increase the IP pool size. IP pools split across all Controllers, so the total pool must be large enough to accommodate this distribution. For both the Controller and Gateway, use the built-in HA mechanisms rather than an external load balancer. The built-in mechanisms are designed to optimize scalability, security, and performance. ![](https://cdn.us.document360.io/e7482ce5-acfa-4c3b-b76c-b6c32472644a/Images/Documentation/external_load_balancer_appgate.png?sv=2026-02-06&spr=https&st=2026-11-21T21%3A21%3A35Z&se=2026-11-21T22%3A21%3A35Z&sr=b&sp=r&sig=UVwzkHjTzbC31S%2BuVD5c8bBjpCb5xJxP6l4nSJEvbQI%3D) #### Accessing Gateways directly using their internal IP addresses There is an optional setting for *Connect via Local Network* in the Gateway’s [Secure Tunnel Settings](/v1/docs/adding-a-gateway-v7-0#secure-tunnel-settings), which allows the Gateway's *Local Client Hostname/IP* (172.17.111.10,11,12) to be added. When the *Use local network* option has been enabled in the [Site](/v1/docs/configuring-sites-v7-0#tunneling), the Controller monitors for clients connecting from the *Public IPs for local network users* list. When a user or device connects to the Controller from one of these public IP addresses, the Controller assumes the user is on the local LAN and issues entitlement tokens that include the *Local Client hostname/IP* values (172.17.111.10, 11, 12) instead of the *Client Hostname/IP* values. Clients access the Gateways on that Site directly, without hairpinning traffic through the firewall. ![](https://cdn.us.document360.io/e7482ce5-acfa-4c3b-b76c-b6c32472644a/Images/Documentation/gateway_direct_local_appgate.png?sv=2026-02-06&spr=https&st=2026-11-21T21%3A21%3A35Z&se=2026-11-21T22%3A21%3A35Z&sr=b&sp=r&sig=OZekaj0h4LB345I2AVcMQEShFdqXgOLsdMvlqxiNLeE%3D) #### Accessing the nearest Geolocated Site Some organizations operate a WAN that links multiple locations, with external access from several defined points. Users in these environments may access different locations depending on their geography. For this scenario, AppGate ZTNA provides the *Override with nearest Site* option. However, with *Disable Source NAT* *on Gateways* enabled, the client IP addresses visible through the connected Site come from the IP Pool (based on IdP settings) and are the same on all Sites. This is not an issue in a pure SDP model, where all Sites are independent and not WAN-connected. This combination introduces two challenges, described below. In the example below, some clients need access to A,B,C, and D via the 112.x Site rather than the usual 111.x Site because of their location. *Assigning Entitlements* By default, entitlements are assigned to one Site only. In this case, A,B,C, and D are assigned to the 111.x Site. When the client connects to the 112.x Site, the entitlements would not be available. There are two ways of overriding this default entitlement assignment when the client must connect to the 112.x Site: 1. Assign a single policy for geolocation use. In the policy, override the entitlements using *Override with nearest Site* under [Override Site](/v1/docs/configuring-policies-v7-0). This assigns A,B,C, and D to the nearest geolocated Site. Enable geolocated Sites by selecting *Use for nearest Site* in the Add/Edit Site page. 2. Use the geolocation claim to assign a policy for the 111.x location or another policy for the 112.x location. Override the entitlements in the latter policy to the 112.x Site using *Override by selecting from list* under [Override Site](/v1/docs/configuring-policies-v7-0)[.](/v1/docs/policies-v6-8) This assigns A,B,C, and D to the 112.x Site instead of the default 111.x Site. *Static routes* Static routing for the client address range (192.168.100.64–254) can be set in only one location on the WAN—in this case, through the 111.x Site. Clients connecting to the 112.x Site still use their 192.168.100.64–254 address, so all return traffic routes to the wrong Site. The *IP Pool mapping* feature in Sites allocates a mapped IP address from the range 192.168.101.1–63 for use on the 112.x Site*.* The *allocation* option reduces the required IP pool size because the number of users on the 112.x Site is much lower than on the 111.x Site. Users always appear to come from the 192.168.101.1–63 address range on the 112.x Site. Set a separate static route for the 192.168.101.1–63 range through the 112.x Site to ensure return traffic routes correctly. *Connect via Local Network* Consider the interaction between *Override with nearest Site* and *Connect via Local Network* when used together. In both cases the Controller modifies entitlement tokens before sending them to the client, so the behavior is predictable. - When a client signs in from a local IP, the Controller applies the standard local IP change and overwrites the Gateway’s public hostname. - The Controller skips the nearest-Site geolocation calculation and assumes the local Site is the nearest one. No further changes are made. ![](https://cdn.us.document360.io/e7482ce5-acfa-4c3b-b76c-b6c32472644a/Images/Documentation/geolocated_appgate.png?sv=2026-02-06&spr=https&st=2026-11-21T21%3A21%3A35Z&se=2026-11-21T22%3A21%3A35Z&sr=b&sp=r&sig=w4LIKqTXZNlvExZuq0ZEfnKcivGvWvzwGzDj%2BenVl68%3D) #### Accessing alternative WAN-conjoined Sites in the event of an outage using a fallback Site Some organizations operate a WAN that links multiple locations, with external access from several defined points. Users typically have access provisioned by location based on their geography. In this example, all clients access A, B, C, and D through the 111.x Site. The 111.x Site has a [fallback Site](/v1/docs/configuring-sites-v7-0#tunneling) specified, in this case the 112.x Site. Two of the clients have *Use fallback Site* enabled in their [policy](/v1/docs/policies-v7-0)—typically the system administrators responsible for restoring the 111.x Site. Only these two clients can fail over to the 112.x Site when the 111.x Site has a problem, preventing the 112.x Site from becoming overloaded. The client controls the fallback process, so an outage affecting the Controller Site does not prevent failover. After a minimum of 30 seconds, the client determines that a Site is unresponsive and switches to the fallback Site. Because users appear on the WAN only once, the same IP address cannot appear in multiple locations. This option works best when *Disable Source NAT on Gateways* is not enabled. In that configuration, each user automatically receives a unique source port per TCP/UDP stream on the fallback Site Gateway IP address, and network routing is already in place. If *Disable Source NAT on Gateways* is enabled, then the *IP Pool mapping* feature in [Sites](/v1/docs/configuring-sites-v7-0#tunneling) should be used so unique routes can be set for each site (see the next example). ![](https://cdn.us.document360.io/e7482ce5-acfa-4c3b-b76c-b6c32472644a/Images/Documentation/wan-conjoined-fallback.png?sv=2026-02-06&spr=https&st=2026-11-21T21%3A21%3A35Z&se=2026-11-21T22%3A21%3A35Z&sr=b&sp=r&sig=PXBKhuKTa4nO09lnPHLWAFoNtvcabS7KyOzSZZCy8JI%3D) #### Accessing WAN-conjoined Sites simultaneously With *Disable Source NAT* *on Gateways* enabled, the client IP addresses visible on the protected network come from the IP Pool (based on IdP settings) and are the same on all Sites. This is not an issue in a pure SDP model, where all Sites are independent and not WAN-connected. For WAN-conjoined Sites, avoid having multiple Gateways respond to ARP requests for the tunnel IP address range (192.168.100.64–254). In this example, users connect to both the 111.x Site and the 112.x Site to access hosts A, B, C, and D. If host C initiates a connection to a user (for example, IP telephony), the router sends an ARP request for the user’s address—say, 192.168.100.100. If the user has the same IP address on both Sites, the router receives two replies. The *IP Pool mapping* feature in [Sites](/v1/docs/configuring-sites-v7-0) performs one-to-one translation of IP addresses (from 192.168.100.64–254 to 192.168.101.64–254) for use on the 112.x Site. The translation option is appropriate here because most users connect to both Sites, so both IP pools must be the same size. The user appears to come from 192.168.100.100 in only one location (the 111.x Site). This configuration also lets you route traffic to a specific Site for a specific purpose—for example, to meet a quality of service (QoS) target for IP telephony. ![](https://cdn.us.document360.io/e7482ce5-acfa-4c3b-b76c-b6c32472644a/Images/Documentation/wan-conjoined-simultaneous.png?sv=2026-02-06&spr=https&st=2026-11-21T21%3A21%3A35Z&se=2026-11-21T22%3A21%3A35Z&sr=b&sp=r&sig=TozoPND5%2BAuh4a%2FShUiwfiFyKGQUUsFBvB%2FYihuq7jo%3D)