Snippets from the feature list, have been pasted below :
* Designed to automatically handle tens of millions of requests per second while maintaining high throughput at ultra low latency
* Network Load Balancer operates at the connection level (Layer 4), routing connections to targets
* Ideal for load balancing of TCP traffic
* Elastic IP support – Network Load Balancer also allows you the option to assign an Elastic IP per Availability Zone (subnet) thereby providing your own fixed IP in a subnet. With full control over your IP addresses, Network Load Balancer can be used in situations where IP addresses need to be hard-coded into DNS records, customer firewall rules, and so forth
* Preserve Source IP address from the client side source IP allowing the back-end to see the IP address of the client and be used by applications for processing – normal firewall rules, including VPC Security Groups, can be used on targets
* High Availability built-in
* Tool to migrate from classic ELB to NLB
* Pricing is based on Load Balancer Capacity Units, or LCUs
* Enhanced Logging – use the Flow Logs feature to record all requests sent to your load balancer
* Long-lived TCP Connections – Network Load Balancer supports long-lived TCP connections that are ideal for WebSocket type of applications that are open for months or years, making them a great fit for IoT, gaming, and messaging applications.
The blog post suggests that you no longer need to pre-warm your Load Balancer in anticipation of a surge in traffic (for example, generated from a TV advert), however i’d strongly recommend testing this first before going into production.