Sendense Documentation

Replication Transport And Routing

How DR replication data moves between sites: the hub-routed path via SHA, the direct site-to-site path, and how routes are chosen.

Documents Home

Concept

Replication Transport And Routing

How DR replication data moves between sites: the hub-routed path via SHA, the direct site-to-site path, and how routes are chosen.

ReadyCurrentreplicationroutingdrnetworking

Control Plane Versus Data Plane

SHA is the control plane on every route. It schedules the sync, selects the appliances, applies the route policy, prepares the recovery-side ingress, and receives health and progress reporting. Changing the route never moves control traffic off SHA.

The route policy changes only the replication data path between the two sites. Both data paths end identically at the target SNA, which passes the stream to the Sendense Controller inside the recovery site.

Roles On The Replication Data Path

Four roles participate in every controller-backed replication sync. SHA is the control plane: it schedules the sync, selects healthy appliances at each site, chooses the data route, and tracks progress and errors. The source SNA reads protected workload data at the protected site. The target SNA is the replication ingress at the recovery site. The Sendense Controller is the per-VM destination appliance that holds the replica disks.

The final hop is the same on every route: the target SNA hands the replication stream to the Sendense Controller inside the recovery site. Controller traffic stays within the recovery site network.

DR replication topology

Both replication routes end the same way: the target SNA receives the replication stream and passes it to the Sendense Controller for that VM inside the recovery site.

DR replication topologyThree zones left to right: protected site, Sendense Hub, recovery site. Protected VMs feed the source SNA. From there a solid hub-routed path runs through SHA to the target SNA, and a dashed direct path runs straight from the source SNA to the target SNA. The target SNA feeds the Sendense Controller. The Controller reports health back to SHA.PROTECTED SITESENDENSE HUBRECOVERY SITESOURCEProtected VMsSource workloadsNODE APPLIANCESource SNAReads workload dataCONTROL PLANESHAOrchestration + relayPER-VM TARGETSendense ControllerPer-VM replica disksNODE APPLIANCETarget SNAReplication ingressRoute ARoute ARoute B: DirectPer-VM ingestHealth
Replication data, solid (hub-routed via SHA)Replication data, dashed (direct site-to-site)Health reporting, dotted (Controller to SHA)

Text equivalent

  • Protected VMs at the protected site are read by the source SNA.
  • Route A, hub-routed: source SNA to SHA to target SNA. Solid lines.
  • Route B, direct: source SNA straight to target SNA. Dashed line.
  • Both routes end at the target SNA, which passes the stream to the Sendense Controller inside the recovery site.
  • The Sendense Controller reports its health to SHA. Dotted line.

Route A: Via SHA

Source SNA to SHA to target SNA to Sendense Controller. Replication data travels inside the always-on encrypted appliance connections, so each site only needs its normal outbound connection to SHA.

Route B: Direct P2P

Source SNA directly to target SNA to Sendense Controller. Replication data crosses the site-to-site network after a live reachability probe. In Automatic mode, a failed probe falls back to Route A.

The Two Data Routes

Hub-routed (Via SHA)
Replication data flows source SNA to SHA to target SNA over the always-on encrypted appliance connections. SHA relays the stream and does not store it. Works wherever both sites can reach SHA; no inbound firewall rules and no site-to-site network are required.
Direct site-to-site (Direct P2P)
Replication data flows straight from the source SNA to the target SNA over your own network path between the sites. Sendense runs a live reachability probe from the source SNA before using this route. Best over private or trusted links such as VPN, MPLS, or a private interconnect.

Route Policy

The route is controlled per replication pattern with the replication route policy, and it is evaluated for each sync.

Automatic
Sendense chooses the best route for each sync. The direct route is used when it is reachable; otherwise the sync automatically uses the hub route. This is the default.
Via SHA
Every sync for the pattern uses the hub route, even if a direct path is available.
Direct P2P
Every sync must use the direct route. If the reachability probe fails, the sync fails with an error rather than silently rerouting through SHA.

Choosing A Route Policy

  • Use Automatic unless there is a specific reason not to. It prefers the direct path when it is reachable and falls back safely to the hub route.
  • Use Via SHA when there is no network path between the sites, when firewall policy only allows outbound connections to SHA, or when replication data must stay inside the Sendense encrypted appliance connections end to end.
  • Use Direct P2P when the sites are connected by a private, high-bandwidth link, replication traffic should stay off the SHA data path, and a hard failure is preferred over silent rerouting.

Why The Target SNA Is On Both Routes

The target SNA is the recovery site's replication ingress on every route, not just the direct one. On the hub route it receives the stream from SHA; on the direct route it receives the stream from the source SNA. In both cases it is the component that writes into the Sendense Controller for that VM.

That is why the recovery-site hop from the target SNA to the Controller has to be planned even in deployments that only ever use Via SHA.

One Destination Listener Per Replica Disk

On the target SNA, Sendense prepares one replication listener for each replica disk of the protected VM and maps each listener to the matching replica-disk export on the Sendense Controller. A three-disk protected VM therefore uses three listeners during its sync.

Listeners are leased from the target appliance's assigned range for the duration of a sync and released afterwards. The size of that range therefore also bounds how many replica disks that appliance can carry concurrently.

Route-Specific Requirements

Requirements differ by route. The first item applies to every deployment; the second applies only when the direct route can be selected; the third applies on every controller-backed route.

SNA to SHA - outbound TCP 443 - always, on every route
A single always-on outbound encrypted connection carrying control, health, backup and hub-routed replication traffic. The hub route needs no inbound connections at either site and no site-to-site network.
Source SNA to target SNA - the target appliance's assigned replication ingress range - Direct P2P, and Automatic when the direct route is selected
The direct route requires that the source SNA can reach the target SNA across the site-to-site network on the replication ports Sendense assigns to that appliance. Reachability is verified with a live probe before each direct-route sync.
Target SNA to Sendense Controller - the Controller's replica-disk replication ports - every controller-backed route
Stays inside the recovery site on the replication network selected in the pattern, and is initiated by the target SNA. Required on the hub route as well as the direct route.

Transport security

The hub route carries replication data inside the encrypted appliance connections end to end. The direct route inherits the security characteristics of the site-to-site network it crosses, so use it over private or trusted links such as a VPN, MPLS circuit or private interconnect. Control and orchestration traffic always flows through SHA on either route.

Multi-SNA Implications

Because SHA resolves a healthy appliance per sync, a site with several SNAs does not have one fixed replication pairing. SHA may select any healthy eligible SNA at the source site and any healthy eligible SNA at the recovery site, and skips appliances that are unhealthy, stale, or in maintenance mode.

For the direct route this means network policy has to cover every permitted source-SNA-to-target-SNA pairing, and each target SNA's own assigned range. Covering only one pairing makes the direct route succeed or fall back depending on which appliances were selected for that sync.

Planning The Network

Replication Network And Firewall Planning is the detailed companion to this page. It has the full connectivity matrix, per-appliance assigned ranges and how to read yours, worked firewall examples for single and multi-SNA sites, security characteristics, a validation checklist and troubleshooting.

What SHA Does For Every Sync

SHA orchestrates every sync regardless of which data route is used.

  • Resolves a healthy SNA at the source site and at the recovery site. Sites can run multiple SNAs; SHA balances work across the healthy ones and skips appliances that are unhealthy, stale, or in maintenance mode.
  • Applies the pattern's replication route policy and, where relevant, the reachability probe result to select the route.
  • Prepares the target-side ingress on the target SNA and confirms the Sendense Controller for the VM is ready to receive data.
  • Starts and monitors the sync, records sync and checkpoint recovery state, and reports progress, RPO status, and errors through the GUI and API.

Scope Of The Routing Choice

  • The route policy controls how replication data travels during sync. Failover behavior is defined by the failover operation, not the route.
  • The direct route is not a separate feature to license or install. It is a per-pattern routing choice.
  • The hub route relays the replication stream through SHA; SHA does not keep a copy.
  • Operators set the policy on the pattern; Sendense applies it to each sync. There is no manual per-sync route selection.

Related Docs