Network Service Tiers is a Google Cloud feature that lets you choose how external traffic reaches your Compute Engine virtual machine instances. The tier you select applies at the resource levelâspecifically on the external IP address of a VMâand determines whether traffic flows through Google's premium global backbone or through regular internet service provider networks. You can mix tiers across different resources in the same project; if you do not specify a tier, Premium Tier is used by default. Internal traffic between VMs within a VPC network always uses Premium Tier infrastructure, regardless of the tier setting on the external IP address.
Both tiers charge for outbound data transfer per GiB delivered, but the rates differ significantly between them. Premium Tier pricing reflects the cost of routing traffic over Google's private fiber network with over 100 points of presence worldwide, which provides lower latency and higher reliability. Standard Tier pricing is lower because traffic routes over public ISP networks before reaching your compute instances, similar to how other cloud providers deliver internet traffic. Neither tier charges for inbound data transfer. The specific per-GiB rates vary based on the destination region and volume tier, with larger volumes receiving lower per-unit pricing on both tiers.
When designing your network architecture, the choice between tiers involves balancing performance requirements against budget constraints. Premium Tier is ideal for applications where low latency and high reliability are criticalâsuch as user-facing services with a global audience or latency-sensitive workloadsâbecause the premium backbone minimizes the number of network hops and avoids congestion on public internet paths. Standard Tier suits workloads where cost optimization matters more than minimal latency, such as batch processing, development environments, or traffic that can tolerate some variability in performance. You can combine both tiers within a single architecture by assigning Premium Tier to latency-sensitive production workloads while using Standard Tier for background jobs or non-critical services, which lets you optimize spend without sacrificing performance where it matters most.
Network Service Tiers are separate from Tier_1 networking performance, which is a per-VM configuration option that provides higher bandwidth between compute instances and carries an additional charge. For egress traffic to the internet, Standard Tier provides the lower pricing tier compared to Premium Tier, and this option becomes attractive when your application does not have strict latency requirements. If you need to minimize costs while maintaining private connectivity, consider using Cloud Interconnect or Cloud VPN for traffic between Google Cloud and on-premises environments, as these dedicated connections can reduce egress charges for high-volume data transfers compared to internet-based routing.
Network Service Tiers in Google Cloud are a resource-level configuration that controls the path external traffic takes to reach cloud resources. The choice between Premium Tier and Standard Tier is a key decision for hybrid and multi-cloud architectures, balancing performance and cost for connectivity to on-premises environments, other clouds, and Google Cloud services.
Premium Tier delivers traffic over Google's private global fiber backbone network, which is designed for high reliability and low latency. External traffic enters this network at one of Google's over 100 global points of presence (PoPs) and is routed internally to the destination VPC network. This path minimizes congestion and the number of hops, providing superior performance and a 99.99% availability SLA for the network path. Premium Tier is the default selection and is mandatory for all internal IP communication within VPC networks, including traffic over VPC Network Peering, Cloud VPN, and Cloud Interconnect. Choose Premium Tier when optimizing for performance and global reach is the primary requirement, such as for users distributed worldwide.
Standard Tier routes external traffic over the public internet using transit ISP networks after it leaves Google's network edge. This path is subject to the performance and reliability characteristics of those intervening networks, offering network quality comparable to other cloud providers. Standard Tier is priced lower than Premium Tier because Google does not carry the traffic on its private backbone for the entire journey. There is no charge for inbound data transfer to Google Cloud for either tier, but outbound pricing per GiB differs. Choose Standard Tier when cost optimization is the primary driver and the performance of the public internet is acceptable for the workload.
The network tier applies to resources with external IP addresses, such as VM instances and external load balancers. For hybrid and multi-cloud architectures, the tier choice directly impacts the performance and cost of traffic flowing between Google Cloud and on-premises data centers or other cloud providers. When connecting to on-premises environments via Cloud VPN or Cloud Interconnect (Dedicated or Partner), the internal IP communication within Google Cloud always uses Premium Tier infrastructure. However, if the on-premises side initiates communication to a Google Cloud resource's external IP address, the selected tier (Premium or Standard) for that external IP dictates the network path for that inbound leg of the journey. For multi-cloud connectivity, especially when using internet-based paths or Cross-Cloud Interconnect, applying Premium Tier to Google Cloud resources ensures the traffic uses Google's backbone up to the interconnect point with the other cloud provider's network, optimizing that segment of the path.
The choice between tiers involves a direct tradeoff between network performance/reliability and cost. Premium Tier should be selected for performance-sensitive frontends, globally distributed user bases, and workloads requiring the highest availability SLAs. Standard Tier is suitable for batch processing, backup traffic, or other scenarios where lower cost outweighs the need for optimal network performance. Architecturally, you can mix tiers within a single project, applying Premium Tier to critical user-facing components like load balancers and Standard Tier to backend services or development environments. This granular control allows for cost-performance optimization tailored to each component's role in a hybrid or multi-cloud application.
Network Service Tiers enable organizations to configure the routing paths and performance profiles of internet-facing traffic connecting to Google Cloud workloads. Google Cloud provides two distinct tiers: Premium Tier, which routes traffic across Google's private global fiber network, and Standard Tier, which routes traffic over public transit internet service providers (ISPs). Network tiers are selected at the resource level, such as on a virtual machine (VM) external IP address or a load balancer forwarding rule, allowing architectures to balance performance needs against data transfer costs. If no tier is specified during resource configuration, Google Cloud applies Premium Tier by default.
Premium Tier delivers external traffic between internet clients and Google Cloud resources by using Google's private, low-latency global backbone network. Inbound traffic from an external client enters Google's network at the point of presence (PoP) closest to that client, and then travels across Google's private fiber directly to the destination VM in the Virtual Private Cloud (VPC) network. Outbound egress traffic follows the reverse path, staying on Google's private backbone across long distances and exiting onto the public internet at the PoP closest to the receiving client. This routing strategy minimizes intermediate ISP hops and network congestion, resulting in low latency, minimal packet loss, and high fault tolerance. Additionally, all internal communication between VM instances using internal IP addresses within VPC networks operates over this high-performance Premium Tier infrastructure.
Standard Tier delivers external traffic by routing data across the public internet and third-party transit provider networks rather than Google's private backbone. For inbound traffic, external packets travel over third-party ISP networks and only enter Google's network at the edge PoP closest to the specific Google Cloud region hosting the destination resource. Outbound egress packets immediately exit Google's network at the PoP nearest to the originating Google Cloud region, leaving the remaining transit across the public internet to reach the destination user. Because packets traverse multiple third-party autonomous systems, network performance and packet loss are subject to the reliability and peering conditions of intervening ISPs. Standard Tier provides network quality and availability comparable to conventional public cloud networking at a lower outbound data transfer rate per delivered gibibyte.
Architectural decisions between Network Service Tiers depend on application latency requirements, user geographic distribution, service compatibility, and budget constraints. Premium Tier is the required choice for global user bases that demand minimal latency, high availability, and support for services like Google Kubernetes Engine (GKE) external Application Load Balancers. Standard Tier is the ideal choice for regional workloads, non-critical data transfers, or cost-sensitive applications where traffic stays within a localized geographic boundary. Network tiers can be configured hierarchically at the project level, cluster level, or overridden at the individual resource level, while inbound data transfer remains free across both tiers. These routing tiers are distinct from per-VM Tier_1 networking, which is a separate compute configuration used to increase raw maximum egress bandwidth for heavy compute workloads.
Premium Tier routes external traffic over Google's private global fiber backbone, entering Google's network at the PoP closest to the client and traveling internally to the destination region. Standard Tier routes traffic over the public internet through third-party ISP networks, only entering Google's network at the PoP closest to the destination region, which results in more hops and potential congestion.
Choose Standard Tier when cost optimization is the primary driver and the workload can tolerate the performance variability of the public internet, such as for batch processing, backup traffic, development environments, or regional workloads where traffic stays within a localized geographic boundary.
Yes, you can mix tiers across different resources in the same project. For example, you can assign Premium Tier to latency-sensitive production load balancers while using Standard Tier for non-critical backend services or development environments, allowing you to optimize cost without sacrificing performance where it matters most.
For hybrid connectivity via Cloud VPN or Cloud Interconnect, internal IP communication within Google Cloud always uses Premium Tier, but traffic initiated from on-premises to a Google Cloud resource's external IP address follows the selected tier. For multi-cloud connectivity, applying Premium Tier to Google Cloud resources ensures traffic uses Google's backbone up to the interconnect point with the other cloud provider, optimizing that segment of the path.
Professional Cloud Network Engineer
Prepare and test your skills
Prepare and test your skills
StudyCerts
StudyCerts