FAQ for site-to-site VPN gateways
You might encounter these frequently asked questions when you use IBM Cloud® VPN for VPC.
What types of VPN connections does IBM Cloud support for VPCs?
IBM Cloud supports the following types of VPNs for VPC:
- Site-to-site VPN: Used for network-to-network connectivity.
- Client-to-site VPN: Used for secure remote access for individual users.
What is IBM Cloud VPN for VPC (site-to-site)?
IBM Cloud VPN for VPC provides secure, encrypted connectivity between your IBM Cloud Virtual Private Cloud (VPC) and other private networks, such as on-premises environments. It uses IPsec tunnels over the internet to protect data in transit. To learn more, see site-to-site VPN gateway and IBM Cloud VPN for VPC Product Page.
What's the difference between VPN for VPC, Transit Gateway, and Direct Link?
The following table provides the difference between VPN for VPC, Transit Gateway, and Direct Link:
| Service | Purpose | Best for |
|---|---|---|
| VPN for VPC | Secure connectivity over the internet | Hybrid cloud, backup links, rapid deployment |
| Transit Gateway | VPC-to-VPC routing | Multi-VPC orchestration within IBM Cloud |
| Direct Link | Dedicated private connections from on premises to VPC | High-bandwidth, low-latency private connectivity |
Why should I use a site-to-site VPN over a client-to-site VPN?
Use a site-to-site VPN to secure data transfer between your IBM Cloud VPC and other private networks, supporting hybrid cloud architectures and inter-network communication. Use a client-to-site VPN when individual users need secure, encrypted access from remote locations, which is ideal for remote work scenarios or third-party contractors.
What types of site-to-site VPN does IBM Cloud VPC support?
IBM Cloud VPC supports the following types of site-to-site VPNs:
- Policy-based VPN: Routes traffic based on defined IP policies, operates in active-standby mode, best for static networks. Learn more.
- Route-based VPN: Uses virtual tunnel interfaces and dynamic routing (BGP), supports active-active redundancy, ideal for complex, scalable architectures. Learn more.
What are typical IBM Cloud site-to-site VPN use cases?
IBM Cloud site-to-site VPN for VPC helps in hybrid cloud extension, backup connectivity, and inter-region VPC communication. See site-to-site VPN use cases.
Can site-to-site VPN for VPC serve as a backup connectivity?
Yes, VPN for VPC is commonly used as a backup to Direct Link or as secondary connectivity alongside Transit Gateway. See High Availability.
How does IBM Cloud site-to-site VPN protect my data?
VPN for VPC uses advanced IPsec/IKEv2 encryption algorithms to ensure confidentiality and integrity of data in transit. Learn more about policy negotiation.
Does the VPN affect my network speed?
VPN encryption adds minimal overhead. IBM Cloud optimizes VPN for reliable performance, but actual speed depends on your network conditions and configuration.
How do I troubleshoot IBM Cloud VPN connection issues?
To troubleshoot IBM Cloud VPN connection issues, check the VPN dashboard tunnel status, validate routes, and review access/firewall rules. See site-to-site troubleshooting.
Are there notifications for VPN service unavailability?
To know about VPN service unavailability, use IBM Cloud standard notification process for maintenance events. See IBM Cloud Status page.
What's the starting point to set up site-to-site VPN for my VPC?
To get started with setting up a site-to-site VPN for your VPC, you must create a VPN gateway. See Creating a VPN Gateway.
How many site-to-site VPN connections can I set up per IBM Cloud VPC?
To determine the number of connections for your site-to-site VPN, see VPN Quotas.
Where can I find pricing details for IBM Cloud site-to-site VPN for VPC?
To find pricing details for IBM Cloud site-to-site VPN for VPC, see the Pricing tab.
What security best practices should I follow with IBM Cloud VPN?
Use strong encryption, update configurations regularly, and apply strict access controls.
When I create a VPN gateway, can I create VPN connections at the same time?
In the IBM Cloud console, you can create the gateway and a connection at the same time. If you use the API or CLI, VPN connections must be created after the VPN gateway is created.
If I delete a VPN gateway with attached VPN connections, what happens to the connections?
The VPN connections are deleted along with the VPN gateway.
Are IKE or IPsec policies deleted if I delete a VPN gateway or VPN connection?
No, IKE and IPsec policies aren't deleted because they can apply to multiple connections.
What happens to a VPN gateway if I try to delete the subnet that the gateway is on?
The subnet can't be deleted if any virtual server instances are present, including the VPN gateway.
Are there default IKE and IPsec policies?
Yes. When you create a VPN connection without referencing a policy ID (IKE or IPsec), auto-negotiation is used.
Why do I need to choose a subnet during VPN gateway provisioning?
You must choose a subnet when you provision a VPN gateway in IBM Cloud because the gateway is deployed within a VPC subnet to establish connectivity. A route-based VPN can support connectivity across all zones, but the gateway itself requires four available private IP addresses in the chosen subnet to provide high availability and automatic maintenance. It is best if you use a size 16 dedicated subnet for the VPN gateway, where the length of the subnet prefix is shorter or equal to 28.
What should I do if I am using ACLs on the subnet that is used to deploy the VPN gateway?
Make sure that ACL rules are in place to allow management traffic and VPN tunnel traffic. For more information, see Configuring network ACLs for use with VPN.
What should I do if I am using ACLs on the subnets that must communicate with an on-premises private network?
Make sure that ACL rules are in place to allow traffic between virtual server instances in your VPC and your on-premises private network. For more information, see Configuring ACLs for use with VPN.
Does VPN for VPC support high-availability configurations?
Yes, VPN for VPC supports high availability in an Active-Standby configuration for policy-based VPNs, and Active-Active configuration for a static, route-based VPN.
Are there plans to support SSL VPN?
No, only IPsec site-to-site is supported.
Are there any caps on throughput for site-to-site VPNaaS?
Up to 650 Mbps of throughput is supported.
Is Pre-Shared Key (PSK) and certificate-based IKE authentication supported for VPNaaS?
Only PSK authentication is supported.
Can you use VPN for VPC as a VPN gateway for your IBM Cloud classic infrastructure?
Yes. The recommended method to connect your classic network to a VPC is to use an IBM Cloud Transit Gateway. See Setting up access to classic infrastructure.
What is a rekey collision between site-to-site VPNs?
A rekey collision occurs when both VPN peers attempt to initiate a rekey at the same time, which can lead to conflicting negotiations, tunnel instability or dropped connections. This issue is commonly observed in IKEv1, because both sides must
use matching key lifetimes, and the protocol lacks collision-handling mechanisms, which makes it unreliable. However, IKEv2 supports asymmetric key lifetimes to gracefully handle simultaneous rekey attempts. If you use IKEv1, rekey collision
deletes the IKE/IPsec security association (SA). To re-create the IKE/IPsec SA, set the connection admin state to down and then up again. To minimize rekey collisions and maintain stable performance, use IKEv2.
How can I send all traffic from the VPC side to the on-premises side in a policy-based VPN?
To send all traffic from the VPC side to the on-premises side, set peer CIDRs to 0.0.0.0/0 when you create a connection.
When a connection is created successfully, the VPN service adds a CIDR 0.0.0.0/0 through the <VPN gateway private IP> route into the default routing table of the VPC. However, this new route can cause routing
issues, such as virtual servers in different subnets not being able to communicate with each other, and VPN gateways not communicating with on-premises VPN gateways.
To troubleshoot routing issues, see Why aren't my VPN gateways or virtual server instances communicating?.
Can I connect my VPN to multiple Transit Gateways in a dynamic route-based VPN connection?
No, each VPN gateway can be connected to only one Transit Gateway.
Is dynamic routing supported when creating a route-based VPN gateway?
Yes, dynamic VPN connection is supported by a route-based VPN gateway. See Creating a VPN gateway to create route-based VPN and select dynamic for the connection type.
What is ASN and why do I need it for dynamic routing?
An Autonomous System Number (ASN) is a unique identifier that is used in Border Gateway Protocol (BGP) to represent an Autonomous System (AS). It functions similarly to a public IP address in an IPsec connection and serves as a key attribute for identifying devices within a network. Each device within the network is assigned to a specific ASN, and without a valid ASN, a VPN gateway cannot successfully establish a BGP session with other devices in the network.
Do I need to attach a Transit Gateway for the dynamic routing connection to work?
Yes, you need to attach a Transit Gateway for the dynamic routing connection to function properly. Without a Transit Gateway, communication between the VPN gateway and your on-premises network doesn't work, even if the IPsec connection is established.
Can I attach an existing Transit Gateway to the route-based VPN for dynamic routing?
Yes, you can attach an existing Transit Gateway to the route-based VPN for dynamic routing. You don't need to create a new Transit Gateway.
Why do I need a Transit Gateway for dynamic route-based connection?
A Transit Gateway is essential for dynamic routing because it acts as a central hub for all connections within your network. The Transit Gateway manages the routing for all spokes, including VPN connections. Without the Transit Gateway the spokes would be unable to communicate with each other, as they rely on the hub Transit Gateway to facilitate and route traffic between them.
What is the use of advertised CIDRs?
Advertised CIDRs are static IP address ranges that are reachable from the VPN and advertised to your on-premises network. These CIDRs are useful for private endpoints that can't be connected directly to a Transit Gateway, such Secrets Manager endpoints or Cloud Object Storage endpoints. By advertising these CIDRs from the IBM VPN to your on-premises network, any resource within your on-premises environment can connect to these endpoints.
Do I need to configure routes when I create a dynamic route-based VPN connection?
No, you don't need to manually configure routes when you create a dynamic route-based VPN connection. All spokes connected to the Transit Gateway automatically connect to your on-premises network.
What is the difference between static and dynamic route-based VPN connection types?
Static routing connection doesn't use BGP for route advertisements, and can't advertise routes to Transit Gateway or on-premises network. For this connection, all routes need to be manually created and managed, whereas for dynamic connection, no manual configuration is required after the initial provisioning and attachment.
How many routes does VPN for VPC support per VPN peer for a dynamic, route-based connection?
Each IBM VPN supports a maximum of 120 routes for each VPN peer in a dynamic routing setup. If this limit is exceeded, the BGP session for that peer automatically shuts down. To restore the session, reduce the number of routes that are advertised from your on-premises peer network to 120 or less, then toggle the connection in the IBM Cloud to reestablish the BGP session.
How many routes does a VPN appliance support for a dynamic, route-based connection?
Each VPN appliance supports a total of 120 routes, regardless of the number of connected peers. If more than 120 routes are received across a combination of peers, the appliance forwards only the first 120 routes to the transit gateway. For example, if two peers each send 70 routes to the VPN appliance, only the first 120 routes are propagated to the transit gateway.
Can I use more than 120 routes for a VPN in a dynamic, route-based connection?
A VPN gateway supports up to 120 routes. If you require more than 120 routes, you can connect the VPN appliance to different on-premises devices. However, keep in mind that this configuration does not guarantee high availability or disaster recovery for the VPN appliance when it is connected to different on-premises devices.
Does IBM complete quarterly ASV scans of data-plane VPN appliances?
Approved Scanning Vendor (ASV) quarterly scanning is a requirement of the Payment Card Industry (PCI) Security Standards Council. ASV scanning of VPN data-plane appliances is solely a customer responsibility. IBM does not use ASVs to scan data-plane appliances because these scans can negatively impact customer workload functions and performance.
What metrics am I charged for if I am using VPN gateway for VPC?
The following metrics are collected for VPN gateway billing on a monthly basis:
- VPN Gateway Instance Hour: How much time your VPN gateway instance is up and running.
- VPN Connection Hour: How much time each of your VPN connections is established and maintained on the VPN gateway.
- Floating IP: The number of active floating IP addresses used by the VPN gateway instance.
When you use a VPN gateway, you are also charged for all outbound public internet traffic that is billed at VPC data rates.
When isn't traffic routed through a route-based VPN gateway?
If you configured a VPC route with a VPN connection as the next hop, traffic might not be routed as expected due to the following conditions:
- The security groups that are associated with the VPC instance don't permit the traffic. In addition, the network ACLs associated with the instance's subnet or the VPN gateway block the traffic. Make sure that your security groups and ACLs allow the intended traffic. For more information, see Configuring network ACLs for use with VPN.
- If the source IP of the traffic doesn't belong to a subnet that is associated with the routing table containing the VPN route, the VPN gateway drops the traffic. For example, consider a VPC routing table that is associated only with Subnet A, and includes a route whose next hop is a VPN connection. If traffic reaches the VPN gateway but originates from an IP address outside of Subnet A or any other subnet that is linked to that routing table, the gateway doesn't forward the traffic.