AWS Interconnect Multicloud with OCI Reaches GA, Replacing DIY Cross-Cloud Networking

What AWS Interconnect – Multicloud Actually Does

The service creates a private, managed network connection between Amazon VPCs and VPCs in other clouds. Customers no longer configure physical or virtual routers, order cross-connects, or manage BGP peering manually. Instead, they select a target cloud provider, a destination region, and required bandwidth. AWS and the provider handle provisioning on both sides.

Every Interconnect is provisioned across redundant network devices spanning at least two physically distinct facilities with independent power and networking. The multicloud offering uses a four-connection model with Equal-Cost Multi-Path load balancing. That means at least one link stays active during planned maintenance.

All physical connections between AWS routers and partner devices are encrypted using MACsec — IEEE 802.1AE MAC Security. The devices transmit customer traffic only when the encryption session is active. Customers see a single logical object in their AWS account. The underlying redundant infrastructure is abstracted away.

Provisioning follows a two-step create-and-accept flow. One side initiates and generates an activation key. The other side uses that key to accept, triggering automated provisioning on both ends. New interconnects are typically provisioned and configured within minutes. Bandwidth can be adjusted after the fact without reprovisioning or opening a support case.

How This Arrived at GA

AWS built the service on an open specification that other cloud providers are adopting. Google Cloud reached GA support before today's announcement. OCI launched support in public preview in May 2026 and is now GA alongside the July 29 announcement. Microsoft Azure is planned to launch later in 2026, but that date is not confirmed in the announcement.

Oracle described the collaboration publicly in April. The two companies planned to establish connectivity between Oracle Interconnect and AWS Interconnect – multicloud, targeting the us-east-1 region. Oracle's announcement noted the connection would support both full and split-stack multicloud deployments. It also stated that customers could avoid managing multiple network providers and physical network infrastructure. Those claims originated with Oracle.

The AWS GA announcement confirms OCI support is live; it does not independently characterize the full scope of Oracle's deployment model.

The service connects to AWS VPCs through Virtual Private Gateways, Transit Gateways, and Cloud WAN. When using Cloud WAN, any Core Network Edge can reach any Interconnect globally that is attached to the same Direct Connect Gateway. Virtual Private Gateways and Transit Gateways can only reach an Interconnect local to their Region — relevant for teams planning cross-region architectures.

All Interconnects include a single CloudWatch Network Synthetic Monitor at no extra cost. That monitor provides round-trip latency and packet loss data. The announcement does not state pricing for the connection itself. Teams should consult the AWS Interconnect pricing page before building cost models.

Why This Matters for Your Architecture

The previous path for AWS-to-OCI connectivity required teams to build and manage global multi-layered networks at scale — provisioning virtual routers, configuring BGP sessions, ordering physical cross-connects, and handling ongoing maintenance. That operational overhead is real and adds lead time.

A fully managed path changes the cost structure and the risk profile at the same time. Managed provisioning removes setup lead time. Built-in resiliency spans redundant devices across physically distinct facilities, reducing single points of failure at the device, cross-connect, and facility level. All physical connections between AWS routers and partner devices are secured using MACsec encryption, designed to ensure data confidentiality and integrity on those physical links.

The announcement does not describe encryption coverage beyond the physical connection boundary.

As Metrotechs analysis: teams running Oracle AI Database workloads on AWS — a scenario Oracle has been actively promoting — are the most immediate candidates for this path. Low-latency, private connectivity between compute on AWS and data services on OCI directly affects query performance and data movement costs. The value scales with data volume and call frequency.

The current GA scope is limited to us-east-1. Teams with workloads in other AWS Regions cannot use the OCI connection today. The announcement does not state when additional regions will become available.

What to Do Before You Commit

Cloud platform owners should take three concrete steps now.

First, inventory your existing OCI-to-AWS connectivity. If your team built a DIY path — virtual routers, third-party SD-WAN overlays, or transit through on-premises equipment — assess whether that path carries production workloads and what SLA it currently delivers. That baseline determines whether a migration is worth the effort.

Second, confirm your workload regions. The OCI connection is available only in us-east-1 today. If your primary AWS workloads run in us-east-2, eu-west-1, or anywhere else, GA availability does not translate to immediate usability. Watch for additional Region announcements before adjusting your roadmap.

Third, review your security and governance controls for cross-cloud traffic. MACsec encrypts physical links between AWS and OCI routers. Security leaders should confirm whether their current policies cover private cloud-to-cloud paths and whether new monitoring, access controls, or logging requirements apply. The announcement does not describe cross-cloud IAM integration or data residency enforcement.

Those details are not stated in the available sources and should be confirmed in the AWS Interconnect documentation before production deployment.

Microsoft Azure support is planned for later in 2026. Teams running three-cloud architectures should treat Azure connectivity as conditional until AWS confirms a GA date.