Skip to main content
Cloud & AI Hub
Browse
Glossary AI Directory Playgrounds Models Prompts Explainers Strategy Matrix Benchmark Decoder

Direct Connect Gateway

An AWS component that lets one Direct Connect private link reach VPCs in many regions and accounts — so you don't need a separate physical connection per region. The hub for private AWS connectivity.

What is a Direct Connect Gateway?

A Direct Connect Gateway is an AWS networking component that lets a single AWS Direct Connect connection reach VPCs across multiple AWS regions and accounts, rather than being tied to one region. Direct Connect itself is AWS’s service for a dedicated, private physical network connection between your on-premises data center and AWS — a private circuit that bypasses the public internet entirely, giving consistent bandwidth, lower and more predictable latency, and enhanced security compared to an internet-based VPN. The Direct Connect Gateway is the piece that makes one such connection flexible: it acts as a hub so your single private link can privately reach resources spread across regions and accounts, without provisioning a separate physical connection for each.

Why it exists — private connectivity beyond the internet

To understand the gateway you need the problem Direct Connect solves and its original limitation. Organizations connecting on-premises infrastructure to AWS have two main options: an IPsec VPN over the public internet (cheap, quick to set up, but rides the unpredictable public internet with no bandwidth or latency guarantees), or Direct Connect — a dedicated private circuit (via an AWS Direct Connect location/partner) that delivers consistent high bandwidth, predictable low latency, and traffic that never touches the public internet — the choice for large, steady data transfer, latency-sensitive workloads, and strict security/compliance requirements. The catch historically was that a Direct Connect connection (via a private virtual interface) attached to a single VPC in a single region. As organizations grew to many VPCs across many regions and accounts, that meant either many expensive physical connections or complex workarounds. The Direct Connect Gateway removes this limit: it’s a globally-available hub that a single Direct Connect connection associates with, and through it that one connection can reach VPCs in any AWS region (except China) and across accounts — so you buy and maintain the physical connectivity once and use it broadly.

Where it fits, and the realities

In practice, the Direct Connect Gateway is often paired with a Transit Gateway for scalable multi-VPC connectivity: the Direct Connect Gateway brings the on-prem private link in, and the Transit Gateway fans it out to many VPCs cleanly (a common modern reference architecture). This whole area — Direct Connect, its gateway, Transit Gateway — is the toolkit for hybrid cloud networking at scale, and it parallels the equivalents on other clouds (Azure ExpressRoute, Google Cloud Interconnect). The realities to weigh: Direct Connect is more expensive and slower to provision than a VPN — you’re arranging a physical cross-connect (port fees, data-transfer charges, and often weeks of lead time via a Direct Connect partner) — so it’s justified by genuine needs for consistent bandwidth, low latency, or private-path compliance, not for small or bursty connectivity where a VPN is cheaper and faster to stand up. Redundancy matters: a single Direct Connect connection is a single point of failure, so best practice is redundant connections (ideally at different locations) and often a VPN as backup — relying on one un-redundant private circuit for critical connectivity is a real risk. And the gateway itself has scope/routing considerations (route limits, no transitive routing between VPCs through the gateway by itself — which is part of why Transit Gateway is used alongside it). The recurring mistakes: choosing Direct Connect (and its cost/lead-time) where a VPN would suffice, deploying a single connection without redundancy or VPN failover (so a circuit failure cuts hybrid connectivity), and misunderstanding the gateway’s routing scope (expecting it to transitively route between VPCs on its own).

What people get wrong

  • Using Direct Connect where a VPN fits — the dedicated circuit’s cost and weeks-long provisioning are justified only by real needs for consistent bandwidth/latency or a private path; a VPN is cheaper and faster for modest needs.
  • No redundancy: a single Direct Connect connection is a single point of failure — critical hybrid connectivity needs redundant circuits and/or a VPN backup, or an outage cuts you off.
  • Misjudging routing scope — the Direct Connect Gateway extends one connection to many regions/accounts but doesn’t by itself transitively route between VPCs; a Transit Gateway is typically paired with it for that.

Primary source: AWS Direct Connect gateways

Historical figures and technical concepts for informational purposes only. Not technical, professional, legal, or financial advice. Sources: Official Documentation.