ExpressRoute Gateway: Scale with Ease

Key insights

• ErGwScale (ExpressRoute Scalable Gateway): A new gateway SKU that lets you set scale units from 1 to 40 and deliver up to 40 Gbps for ExpressRoute connections.
  It enables dynamic bandwidth scaling to match workload peaks and troughs.
• Deployment and configuration: Choose Gateway type = ExpressRoute and SKU = ErGwScale, then set Minimum and Maximum Scale Units (1–40).
  Changes take about 30 minutes and gateways are zone‑redundant by default.
• Scaling modes: Use a fixed-size gateway by setting minimum = maximum, or enable autoscaling by setting minimum ≥ 2 and a higher maximum.
  Autoscaling adjusts capacity automatically as demand changes.
• Key limits and capacity: Circuit limits scale with units (4 circuits at 1 unit, 8 at 2 units, 16 at 10+ units).
  VM connection capacity scales at 2,000 VMs per unit for the first 10 units, then 1,000 per unit beyond.
• Features and interoperability: Supports ExpressRoute FastPath (requires minimum 10 units), allows VPN and ExpressRoute coexistence, and works with Global Reach and zone‑redundant deployments for higher resilience.
• Operational benefits and considerations: Offers cost efficiency by scaling only when needed and enables no‑downtime upgrades from older ErGw SKUs via portal or PowerShell.
  Plan for regional availability and pricing that depend on chosen scale units.

Introduction

In a new YouTube video, John Savill's [MVP] walks viewers through Azure's updated virtual network gateway for ExpressRoute, explaining how the new ErGwScale SKU introduces flexible, high-bandwidth connectivity. The video breaks the topic into clear chapters, covering gateway basics, SKU differences, scaling behavior, limits, pricing, and migration paths. Consequently, the presentation gives IT teams a practical view of when and how to adopt the scalable gateway. Moreover, the host balances technical detail with real-world scenarios to help network architects weigh options.

Gateway Basics and Deployment

John begins by defining the core role of an ExpressRoute gateway: it links on-premises networks to Azure virtual networks over private circuits rather than the public internet. He highlights that the deployment process follows familiar steps in the Azure portal, but now includes parameters for minimum and maximum scale units, which control the gateway's throughput and capacity. In addition, he notes that the scalable gateway is zone-redundant by default and uses Microsoft-managed public IPs for policy and auditability. Thus, teams can use standard deployment workflows while gaining new controls for performance.

SKU Types, Scale Units, and How Scaling Works

The video contrasts the new ErGwScale SKU with legacy fixed SKUs such as ErGw1Az, ErGw2Az, and ErGw3Az, emphasizing that ErGwScale lets you set a range from one to forty scale units. For fixed-size deployments, minimum and maximum units match; for autoscaling, the minimum is set to two or more while the maximum can go up to forty. John then explains the mechanics: autoscaling increases capacity based on observed demand and back-end provisioning, and typical configuration changes complete in under 30 minutes. Therefore, teams can expand bandwidth to as much as 40 Gbps without scheduled downtime during upgrades.

He also addresses practical limits and capacity planning: some features require minimum unit counts, for example, FastPath needs at least ten units to enable path bypass for certain TCP/UDP flows, which lowers latency. Additionally, the number of supported ExpressRoute circuits and VM connection counts scale with units—small deployments get fewer circuits and lower VM capacity, while larger unit counts enable up to 16 circuits and thousands of VM connections. As a result, architects must match scale unit choices to circuit needs and VM density to avoid bottlenecks. In short, the SKU provides granularity, but it also requires deliberate planning to align with workload topology.

Pricing, Migration and Operational Tradeoffs

John covers pricing implications plainly: autoscaling can reduce over-provisioning costs by matching capacity to demand, yet higher maximum unit settings will increase potential peak costs. Therefore, financial tradeoffs emerge between setting conservative fixed sizes and allowing autoscaling with a higher cap. He recommends that organizations run usage profiling to choose sensible minimums and maximums, and to combine scale planning with budget governance to avoid unexpected charges.

Regarding migration, the video explains that upgrades from fixed SKUs to ErGwScale can be performed through the portal or via scripting tools and typically complete without service interruption. However, he warns that not all regions support the scalable SKU yet, and that careful verification of regional availability is essential before planning migration. Consequently, teams must coordinate migrations with regional compliance and availability requirements to ensure a smooth cutover. In practice, migration is straightforward technically but requires operational checks.

Challenges and Considerations for Adoption

John discusses several challenges that teams face when adopting the scalable gateway, including monitoring complexity and the need to set sensible autoscale thresholds. Autoscaling reduces manual intervention, but it also introduces variability in capacity that can make cost forecasting harder. Moreover, some features like FastPath demand higher unit counts, so enabling performance optimizations sometimes forces organizations to accept higher baseline costs. Thus, decision-makers must balance latency, throughput, and budget goals.

He also highlights limits and regional exclusions that complicate universal adoption. For example, certain Azure regions do not yet support the scalable SKU, and VM connection ratios change after a threshold of units, which can affect large-scale migration plans. In addition, coexisting VPN and ExpressRoute configurations, and the need for Global Reach or zone-redundant patterns, add design choices that require careful testing. Therefore, the path to modernizing network gateways is less about a single switch and more about staged validation and governance.

Outlook and Recommendation

In closing, John projects that the ErGwScale SKU represents a meaningful step toward more cloud-native, demand-driven networking in Azure, and he suggests that organizations with variable or growing throughput needs should evaluate it quickly. He also recommends testing autoscaling in a controlled environment to observe real scaling behavior and to refine minimum and maximum settings based on actual traffic. Finally, he urges teams to weigh immediate performance benefits against longer-term cost predictability, since the right choice depends on workload patterns and organizational priorities.

Keywords

Azure ExpressRoute Gateway, ExpressRoute scalable gateway, ExpressRoute gateway scalability, ExpressRoute gateway autoscale, ExpressRoute gateway performance, ExpressRoute gateway SKUs, ExpressRoute high availability, ExpressRoute gateway best practices