AWS Elastic Load Balancing

Overview

ELB combines four load balancer types with automatic scaling of request handling and configurable target health checks. That makes it a practical choice for workloads whose traffic varies and teams that want managed load balancing. Classic, Application, and Network Load Balancers carry a monthly availability guarantee of at least 99.99%; that guarantee does not extend to Gateway Load Balancer in the stated terms.

One meaningful migration consideration is API compatibility: Application Load Balancers use a different API set from Classic Load Balancers. Teams carrying forward Classic-based tooling should account for that distinction before switching.

Key features

• Workload-specific routing: Application Load Balancer is the recommended fit for HTTP requests, with header-based routing, redirects, fixed responses, HTTP/2, and gRPC. Network Load Balancer is geared to transport protocols and workloads where extreme performance or low latency is a priority. Gateway Load Balancer suits third-party virtual appliances, while Classic Load Balancer targets applications built in EC2-Classic.
• Security and visibility: SSL/TLS termination, certificate management, and client certificate authentication cover key connection-security needs. CloudWatch metrics and logging are available across all four types, giving teams a consistent monitoring integration.
• Network and hybrid options: Network Load Balancer supports static Elastic IP addresses and AWS PrivateLink for TCP and TLS listeners. A single ELB load balancer can also distribute traffic across AWS and on-premises resources, useful for hybrid deployments.
• Kubernetes: Controller support includes direct-to-pod routing for Application Load Balancers and support for fully private EKS clusters with Application and Network Load Balancers. These options make ELB relevant to Kubernetes workloads that need those specific integrations.

Pricing

ELB is paid, with usage-based charges that vary by region. There is no free trial. The free tier can offset some early usage, but its shared hours and processing allowances are bounded, and it lasts only six months after account creation.

• AWS Free Tier — 0.00 USD per free: New AWS customers receive 750 hours per month shared between Classic and Application Load Balancers, plus 15 GB of Classic Load Balancer data processing or 15 LCUs for Application Load Balancers. The plan also includes up to $200 in credits and is available for six months after account creation. It is a limited starting allowance, not a general free tier for every ELB type.
• Pay-as-you-go — Gateway Load Balancer — custom pricing: Charged per hour or partial hour running, plus Gateway Load Balancer Capacity Units used per minute; rates vary by region. Gateway Load Balancer Endpoint is billed separately, and data transfer and public IPv4 charges may also apply. This structure suits usage-based appliance routing, but endpoint and ancillary costs need to be included in estimates.
• Pay-as-you-go — Application Load Balancer — custom pricing: Charged per hour or partial hour running, plus Load Balancer Capacity Units used per minute, with regional rate variation. Data transfer and public IPv4 charges may be additional. This fits HTTP workloads that need its routing capabilities, though the free-tier allowance is limited to 15 LCUs.
• Pay-as-you-go — Classic Load Balancer — custom pricing: Charged per hour or partial hour running and per GB transferred, at region-dependent rates. Data transfer and public IPv4 charges may also apply. It is the targeted option for EC2-Classic applications, but usage-based traffic charges make volume relevant to budgeting.
• Pay-as-you-go — Network Load Balancer — custom pricing: Charged per hour or partial hour running, plus Network Load Balancer Capacity Units used per minute; rates vary by region. Data transfer and public IPv4 charges may be extra. This is the option for transport-protocol and low-latency needs, with costs tied to both running time and capacity use.

Standard AWS data transfer and public IPv4 address charges may apply in addition to load balancer charges. Gateway Load Balancer Endpoint pricing is separate as well, so the base load balancer charge alone may not represent total cost.

Platforms

ELB is available through an API and the web. Its managed-service model and AWS, hybrid, and Kubernetes integrations make it most relevant to teams already working with those environments.

Who it's for

Choose ELB when you need managed traffic distribution within AWS, across AWS and on-premises resources, or for supported Kubernetes deployments. Its distinct Application, Network, and Gateway options are useful when HTTP features, transport performance, or virtual-appliance routing call for different approaches. It is less compelling for teams seeking a flat, predictable price, or for Classic-based environments where the Application Load Balancer API change would create migration work.

Pros and cons

• Pro: Four load balancer types let teams align routing to HTTP, transport, appliance, or EC2-Classic workloads rather than use one approach for all.
• Pro: Automatic capacity scaling and configurable health checks support changing traffic and target health without operating the load balancer directly.
• Pro: CloudWatch metrics and logging cover all four types, while security and hybrid capabilities address needs beyond basic traffic distribution.
• Con: Hourly, capacity-unit, data-processing, or per-GB charges vary by type and region; data transfer and public IPv4 charges can add to the bill.
• Con: The free tier is time-limited and its 750 hours are shared between Classic and Application Load Balancers, so it is not a lasting cost substitute.
• Con: Application and Classic Load Balancers have different API sets, which can complicate continuity for Classic integrations.

Alternatives

For a broader comparison, see Load Balancing Software. VMware Workstation Pro is a free option for 64-bit Intel or AMD Windows and Linux hosts, with no subscription or license key required. HAProxy Community Edition is a free Linux or self-hosted choice for TCP and HTTP proxying. MetalLB is a free open-source project for Kubernetes load balancing on Linux or self-hosted environments.

F5 BIG-IP APM is a paid alternative with a free trial and a perpetual-license option for BIG-IP Virtual Editions, hardware, or hybrid environments. OpenELB, kube-vip, and PureLB are free Linux or self-hosted alternatives. Hetzner Cloud Volumes is a paid web-based alternative.

Verdict

ELB is a strong choice for teams running AWS-centered applications that need managed scaling, health checks, and a load balancer type suited to their traffic. Its workload-specific options and integrations are the main reasons to choose it; usage-dependent charges and the Application-versus-Classic API difference are the clearest reasons to look elsewhere or plan carefully.