Elastic Load Balance
Flexible Engine / Network
Enhance your performance by automatizing you load balancer
The ability to establish a Load Balancer between two DataCenters from same availability area allows you to easily implement high availability strategies.
A Load Balancer allows up to 100,000 concurrent connections, and up to 5 auditors. An auditor can manage up to 100 virtual servers in back end mode.
The service is invoiced on a per-use basis.
The pay-per-use load balancer allows network traffic to be distributed between virtual servers grouped in pools.
Your Load Balancer can distribute visible loads from the Internet or environments entirely internal of your environment.
This service is an essential tool to set up high availability architectures. For example, it is possible to simply use a load balancer to operate servers in multi-master mode, each contributing to processing part of the traffic.
The load balancer can be installed in a few very simple steps via the console or via APIs. It can distribute any type of TCP connection as well as http and HTTPS sessions.
This service is particularly suitable for the implementation of multi-site architecture since it allows traffic to be distributed between different available zones (for example between the two servers based in France).
– Load balancing on HTTP / HTTPS / TCP flows to a pool of servers
– Supports SSL terminations
– Inter-AZ load balancing
– Autoscaling according to traffic
– Linear expansion capacity (elimination of SPOFs)
– Supports metric monitoring: incoming and outgoing traffic, new requests, simultaneous requests, incoming and outgoing data packets, number of connections, inactive connections.
– Configuration of Health Check on instances
Load distributor for a private network
The private network load balancer has only private IP addresses. It can only direct requests from clients with access to the VPC of the load balancer.
If your application has several levels, for example if you have web servers that must be connected to the Internet and database servers that are only connected to web servers, you can design an architecture that uses internal and web load balancers.
Create a public network load balancer and register web servers with it. Create a private network load balancer and register the database servers with it. Web servers receive requests from the public network load balancer and send requests for database servers to the public network load balancer. Database servers receive requests from the public network load balancer.
The private network load balancer sends requests to different available areas.
Application scenario: Disaster Recovery
ELB distributes incoming traffic across AZs, facilitating real-time disaster recovery. Enterprises like banks can use ELB to ensure high service availability.