This topic is for network engineers. It explains how to configure the on-premises device (the customer-premises equipment, or CPE) at your end of the IPSec VPN so traffic can flow between your on-premises network and virtual cloud network (VCN). See these related topics:
- Overview of Networking: For general information about the parts of a VCN
- VPN Connect: For various topics about IPSec VPNs
- Verified CPE Devices: For a list of CPE devices Oracle has verified
The following figure shows the basic layout of the IPSec VPN connection.
There are several requirements and prerequisites to be aware of before moving forward.
For important details about routing for your IPSec VPN see Routing for the Oracle IPSec VPN.
Oracle uses asymmetric routing across the multiple tunnels that make up the IPSec VPN connection. Even if you configure one tunnel as primary and another as backup, traffic from your VCN to your on-premises network can use any tunnel that is "up" on your device. Configure your firewalls accordingly. Otherwise, ping tests or application traffic across the connection will not reliably work.
If you use BGP dynamic routing with your IPSec VPN, you can configure routing so that Oracle prefers one tunnel over the other.
Note that the Cisco ASA policy-based configuration uses a single tunnel.
Creation of Cloud Network Components
You or someone in your organization must have already used the Oracle Console to create a VCN and an IPSec connection, which consists of multiple IPSec tunnels for redundancy. You must gather the following information about those components:
- VCN OCID: The VCN OCID is a unique Oracle Cloud Infrastructure identifier that has a UUID at the end. You can use this UUID or any other string that helps you identify this VCN in the device configuration and doesn't conflict with other object-group or access-list names.
- VCN CIDR
- VCN CIDR subnet mask
For each IPSec tunnel:
- The IP address of the Oracle IPSec tunnel endpoint (the VPN headend)
- The shared secret
You also need some basic information about the inside and outside interfaces of your on-premises device (your CPE). For a list of the required information for your particular CPE, see the links in this list: Verified CPE Devices.
Oracle recommends that you disable NAT-T at your CPE when establishing IPSec tunnels with Oracle Cloud Infrastructure. Unless you have multiple CPEs sharing the same NAT IP, NAT-T is not required.
If your CPE is behind a NAT device, you can provide Oracle with your CPE's IKE identifier. For more information, see If Your CPE Is Behind a NAT Device.
Route-Based Versus Policy-Based IPSec
The Oracle VPN headends use route-based tunnels, but can work with policy-based tunnels with some caveats listed in the following section.
Oracle supports only a single encryption domain (also known as a proxy ID, security parameter index (SPI), or traffic selector).
For more information, see Supported Encryption Domain or Proxy ID.
IPSec VPN Best Practices
- Configure all tunnels for every IPSec connection: Oracle deploys multiple IPSec headends for all your connections to provide high availability for your mission-critical workloads. Configuring all the available tunnels is a key part of the "Design for Failure" philosophy. (Exception: Cisco ASA policy-based configuration, which uses a single tunnel.)
- Have redundant CPEs in your on-premises locations: Each of your sites that connects with IPSec to Oracle Cloud Infrastructure should have redundant CPE devices. You add each CPE to the Oracle Cloud Infrastructure Console and create a separate IPSec connection between your dynamic routing gateway (DRG) and each CPE. For each IPSec connection, Oracle provisions two tunnels on geographically redundant IPSec headends. Oracle may use any tunnel that is "up" to send traffic back to your on-premises network. For more information, see Routing for the Oracle IPSec VPN.
Consider backup aggregate routes: If you have multiple sites connected via IPSec VPNs to Oracle Cloud Infrastructure, and those sites are connected to your on-premises backbone routers, consider configuring your IPSec connection routes with both the local site aggregate route as well as a default route.
Note that the DRG routes learned from the IPSec connections are only used by traffic you route from your VCN to your DRG. The default route will only be used by traffic sent to your DRG whose destination IP address does not match the more specific routes of any of your tunnels.
Confirming the Status of the Connection
After you configure the IPSec connection, you can test the connection by launching an instance into the VCN and then pinging it from your on-premises network. For information about launching an instance, see Launching an Instance. To ping the instance, the VCN's security rules must allow ping traffic.
You can get the status of the IPSec tunnels in the API or Console. For instructions, see To view the status and configuration information for the IPSec tunnels.
For links to the specific configuration information for each verified CPE device, see Verified CPE Devices.