Oracle adds FastConnect traffic draining for planned maintenance

Oracle added traffic draining for private FastConnect virtual circuits in a release dated 16 September 2026. For teams maintaining hybrid connections to Oracle Cloud Infrastructure, the change offers a way to steer traffic away from a circuit before planned work. The release notice covers FastConnect partner connections, FastConnect direct and Oracle Interconnects.

What changed

The same release adds minimum-links, interface hold-timer and Letter of Authorization enhancements for FastConnect direct cross-connects and cross-connect groups. These features have different scopes: the direct-connection controls should not be assumed to apply to every partner circuit.

Oracle's FastConnect overview explains that draining deprioritizes route advertisements sent toward customer-premises equipment using the Border Gateway Protocol (BGP) graceful-shutdown community 65535:0 together with autonomous-system (AS) path prepending. Oracle recommends this facility for customer maintenance, such as replacing an edge router or upgrading its software.

Minimum-links settings determine how many member connections must remain operational for a cross-connect group to be treated as up. The interface hold timer delays signaling an interface-down condition, helping damp brief flaps. Those controls deserve separate evaluation because capacity loss and a short link interruption present different operational problems.

Why it matters for BYOIP users

The connection to BYOIP is operational: organizations using their own public addresses may also depend on private hybrid links for application backends, migration traffic or management access. A maintenance event on those links can affect the service behind an otherwise unchanged public endpoint.

This release concerns private-circuit maintenance. It does not announce a change to public-prefix import eligibility or Internet advertisement. Keep those responsibilities separate when reviewing the OCI BYOIP integration.

Our editorial recommendation is to add a deliberate traffic-shift stage to the maintenance runbook where the network design supports it. Record the expected route and traffic state before taking a circuit offline, and define the conditions for abandoning the maintenance attempt.

What to check

• Verify that the alternative path has enough capacity and reaches the required destinations. A second circuit should be examined for shared physical dependencies.
• Review how your edge devices and connectivity provider handle graceful-shutdown communities and AS-path changes. Existing routing preferences can affect the intended shift.
• During an approved test window, observe route selection and traffic counters before progressing to circuit shutdown. Include application checks and the return to the preferred path.
• Review stateful firewalls for asymmetric-routing behavior during the transition. Oracle's maintenance guidance specifically calls out this risk and the need to validate failover.
• Choose minimum-links and hold-timer settings against your capacity requirements and failure-detection objectives, rather than copying values from another deployment.

Sources and further reading

The dated release establishes availability. The linked overview and maintenance guide describe behavior and prerequisites. This article provides a planning checklist; it does not report hands-on testing or guarantee an interruption-free change.