AWS Elastic Load Balancing

AWS Elastic Load Balancing (ELB) distributes application traffic across multiple targets and virtual appliances in one or more Availability Zones. It provides Application, Network, Gateway, and Classic Load Balancers, scaling request capacity as traffic changes and running configurable health checks on targets. Application Load Balancers support HTTP header routing, redirects, fixed responses, HTTP/2, and gRPC. Network Load Balancers support static Elastic IP addresses and AWS PrivateLink for TCP and TLS listeners. ELB also offers SSL/TLS termination, certificate management, and client certificate authentication. CloudWatch metrics and logging cover all four types. Kubernetes features include direct-to-pod routing for Application Load Balancers and support for fully private EKS clusters with Application and Network Load Balancers. A single load balancer can serve AWS and on-premises resources. AWS gives guidance on choosing a type based on workload and protocol. Classic, Application, and Network Load Balancers carry a stated monthly availability guarantee of at least 99.99%. The managed service is available through API and web. Charges are usage based, with possible additional data transfer and public IPv4 address costs. New AWS customers receive a six-month Free Tier allowance.

ELB offers several managed load balancer types, with routing, health checks, and monitoring features. Check the usage-based charges and API requirements for the load balancer type you need.

It suits teams distributing application traffic across targets, including those using AWS and on-premises resources or Kubernetes. The four load balancer types address different protocols and workloads.

Overview

AWS Elastic Load Balancing (ELB) is a managed AWS service for distributing application traffic across targets in one or more Availability Zones. It best suits teams that want traffic handling to scale with demand without managing the load balancer themselves. Its range of load balancer types is a strength, but the usage-based bill and distinct API requirements make workload fit and cost planning important.

Key features

ELB scales request-handling capacity automatically and runs configurable health checks on targets. That combination is useful for AWS applications with changing traffic, though it does not remove the need to select the right load balancer type for the traffic involved.

Application Load Balancer handles HTTP requests with header-based routing, redirects, fixed responses, HTTP/2, and gRPC. Network Load Balancer is the option for transport protocols and workloads prioritizing extreme performance or low latency; static Elastic IP addresses and AWS PrivateLink support TCP and TLS listeners. Gateway Load Balancer is aimed at third-party virtual appliances, while Classic Load Balancer is intended for applications built in EC2-Classic.

ELB also supports SSL/TLS termination, certificate management, and client certificate authentication. CloudWatch metrics and logging are available across all four types. Kubernetes Controller support includes direct-to-pod routing for Application Load Balancers and fully private EKS clusters for Application and Network Load Balancers. A single load balancer can also serve AWS and on-premises resources, which is useful for hybrid setups.

AWS recommends Application Load Balancer for HTTP, Network Load Balancer for transport protocols and performance-sensitive or low-latency workloads, Gateway Load Balancer for virtual appliances, and Classic Load Balancer for EC2-Classic applications. Classic, Application, and Network Load Balancers carry a monthly availability guarantee of at least 99.99%. One migration consideration is that Application and Classic Load Balancers use different API sets, so tooling built for Classic may need adjustment.

Pricing

ELB is paid, with a free plan and no free trial. Charges depend on usage, and standard AWS data-transfer and public IPv4 address charges may add to the bill. Rates vary by region, so the operational model is flexible but costs are less predictable than a fixed subscription.

• AWS Free Tier: free (billed). New AWS customers receive 750 hours per month shared between Classic and Application Load Balancers, plus 15 GB of Classic data processing or 15 Application Load Balancer LCUs. The free plan lasts six months after account creation and includes up to $200 in credits. The shared hours and capped processing or LCU allowances make this a starting allowance, not an unlimited production tier.
• Pay-as-you-go — Application Load Balancer: custom pricing, charged per running hour or partial hour and per LCU used per minute. Data transfer and public IPv4 charges may apply.
• Pay-as-you-go — Network Load Balancer: custom pricing, charged per running hour or partial hour and per Network Load Balancer Capacity Unit used per minute. Data transfer and public IPv4 charges may apply.
• Pay-as-you-go — Gateway Load Balancer: custom pricing, charged per running hour or partial hour and per Gateway Load Balancer Capacity Unit used per minute. The Gateway Load Balancer Endpoint is billed separately; data transfer and public IPv4 charges may also apply.
• Pay-as-you-go — Classic Load Balancer: custom pricing, charged per running hour or partial hour and for each GB transferred. Data transfer and public IPv4 charges may also apply.

These plans suit teams that want charges tied to operation and load rather than a fixed tier. The tradeoff is that capacity-related units, runtime, regional rates, and possible network or address charges all matter to forecasting.

Platforms

ELB supports API and web access. Its managed-service model and AWS integrations make it a natural fit for AWS environments; teams seeking a standalone or self-hosted load balancer should compare alternatives.

Who it's for

ELB is a strong choice for AWS teams needing managed Layer 4 or Layer 7 traffic distribution, health checks, TLS termination, or hybrid balancing across cloud and on-premises resources. Its separate types address HTTP, transport, and virtual-appliance workloads, while CloudWatch and Kubernetes Controller support help in environments already using those AWS services.

It is less compelling for teams that need a predictable flat price, want a free ongoing tier, or depend on Classic Load Balancer APIs while moving to Application Load Balancer. EC2-Classic users may still have a specific reason to choose Classic, but the API distinction matters when changing types.

Pros and cons

• Pro: Four load balancer types let teams match HTTP, transport, appliance, or EC2-Classic traffic to different routing needs.
• Pro: Automatic scaling and configurable health checks reduce the need to manage traffic capacity and target health manually.
• Pro: CloudWatch metrics and logging span all four types, with Kubernetes and hybrid capabilities for relevant deployments.
• Con: Usage-based charges vary by region and can include separate data-transfer, public IPv4, or endpoint costs, complicating estimates.
• Con: The free allowance is limited to six months and has shared-hour and processing or LCU caps.
• Con: Application Load Balancer's API differs from Classic, which can create migration work for existing integrations.

Alternatives

For a broader comparison, browse the Load Balancing Software list.

• VMware Workstation Pro is free at no cost, with no subscription or license key required, for supported Windows or Linux hosts on 64-bit Intel or AMD PCs.
• HAProxy Community Edition is a free Linux and self-hosted option for TCP and HTTP proxying.
• MetalLB is a free open-source project for Kubernetes load balancing, with Linux and self-hosted platforms.
• Array APV is another paid option with API, Linux, self-hosted, and web platforms.
• OpenELB is a free option with Linux and self-hosted platforms.
• Loadbalancer.org Enterprise is a paid option with a free trial and API, Linux, self-hosted, and web platforms. Its Enterprise VA Basic perpetual license costs $2,360, paid once and includes 400Mbps throughput, 400 SSL TPS, and unlimited servers; a $150 /mo subscription is also offered.

Verdict

Choose ELB if your workloads run on AWS and you want managed, scalable traffic distribution with a choice of routing types, health checks, and CloudWatch visibility. Its strongest case is the breadth of AWS-integrated options, including Kubernetes and hybrid use. Look elsewhere if predictable fixed pricing or a different deployment model matters more, and account for API compatibility before moving from Classic to Application Load Balancer.