DVR vs Non-DVR

This articles describes the two ways to configure networking in PMO.

Networking in PMO can be configured in two modes:

  1. Distributed Virtual Routing (DVR / L3-DVR) or,
  2. Legacy mode.

When setting up networking for your PMO cloud, figuring out if you will need to setup Distributed Layer 3 (L3) versus centralized L3 control is important. The mode that you choose will depend on your specific use case, with each mode having it’s own benefits and drawbacks.

DVR configuration has following characteristics:

  • Distributed Layer 3 (L3) space across all compute hosts. No dedicated network nodes.
  • Requires dedicated node/nodes for L3 Source Network Address Translation (SNAT).
  • North-South and East-West traffic is handled at each compute host.

Advantages of DVR

  • High performance (L3 at compute node)
  • Single hop Floating IP (FIP) i.e. North-South traffic
  • East-West traffic also is routed at source.
  • Failure domain reduced to a single compute node.
  • Great load sharing in terms of L3 routing & DHCP servers.
  • Easier for maintenance when it comes to re-purposing the otherwise central network node.

When is DVR not a Good Fit?

  • Needs external network for all compute nodes.
  • Exhausts one additional FIP per compute node.

DVR Vs Non-DVR Deployment

Components in a DVR deployment are distributed as shown below.

Neutron DVR

The Network components in the above picture consist of the dhcp-agent and the l3-agent. DHCP agent is responsible for managing DHCP servers for tenant networks along with DNS service while the L3 agent is responsible for routing i.e. cross networks both east-west and north-south traffic. You can assign one or more nodes to run SNAT. VMs without Floating IPs will connect external to the cloud through this SNAT service node. VMs with floating IPs however will directly communicate external to the cloud from the host/server it is running in.

In comparison, components in a non-DVR deployment is architected as below:

Legacy Non-DVR Neutron deployment

One or more network nodes are dedicated to run the network services. Compute nodes primarily run compute workloads. All cross network traffic (east-west & north-south) will be routed by the routers present in the network node. For example if two VMs on the same server but on different networks ( and need to talk to each other, they would need to go through one of the network nodes. In case of DVR, this routing will take place on the same server.

A condensed side-by-side comparison of the two architectures/setups is given below.

Feature Feature detail Non-DVR Centralized Routing - L3 Distributed Virtual Routing - L3
DHCP server Setup On network node. Redundancy limited to number of network nodes. On any hypervisor. User has the control to choose which hypervisors host DHCP servers. This is scalable to 'N' where N is the number of hypervisors
Failure Domain Active-Active if no. of network nodes > 1 Active-Active
Failover Moves to other network node if network node > 2. Moves to another hypervisor/node with DHCP service enabled if they are > 2
L3 or IP Routing Setup On network node only. No active-active redundancy. On all hypervisors for East West traffic and on SNAT enabled hypervisors for North South traffic.
SNAT On a particular network node. Load balanced if network node > 2 On a particular hypervisor. Load balanced to the hypervisors enabled with SNAT
Floating / Elastic IPs On the same network node as SNAT On the specific hypervisor where VM runs.
North-South traffic flow Through network node. Both SNAT & FIP SNAT through SNAT hypervisor but when using FIP, directly through hypervisor running the VM.
East-West traffic flow Through network node Directly at hypervisor.
redundancy active-active possible (In our roadmap) Active-Active only makes sense for SNAT. (In our roadmap)
Failure Domain Loss of L3 connectivity till failover. Loss of hypervisor alone. If hypervisor running SNAT, loss of SNAT till failover.
Failover L3 routers (SNAT & FIP) failover to to other routers if network_nodes > 1 after timeout. SNAT router failover to other hypervisor after timeout.
IP usage external IPs 1 for external n/w gateway interface per L3 virtual router. 1 each for FIP 1 for external n/w gateway interface per L3 virtual router (SNAT), 1 each for FIP gateway/namespace (=no. of hypervisors) + 1 FIP per VM.
Internal IPs 1 per tenant network per router 2 per tenant network per router ( 1 for SNAT/ 1 for Gateway)