This article is more than 1 year old

F5 hammers out a virtual load balancer

LineRate Point brings virtualisation to old enterprise apps

After two years digesting its acquisition of LineRate, F5 Networks is now shipping one of the key outcomes of the buy, a fully-virtualised lightweight load balancing solution.

As part of a broader virtualisation push that also includes an iron-free version of its Big-IP application delivery controller.

According to director of business development Cyrus Rafii, the new LineRate Point launch is pitched not at the largest-scale Web-facing applications, but to give more run-of-the-mill enterprise applications similar load-balancing and availability benefits.

While having a relatively dull departmental-level application go dark isn't the business ruin than losing a Web booking system, Rafii said, it still disrupts the business – and there's lots of those applications around.

Hence, he explained, LineRate Point fits in those departmental applications that are “used by tens or a hundred people, but there may be hundreds or thousand such applications in an enterprise.”

“That's what makes this new territory for F5 – it completes a portfolio of application delivery,” Rafii said.

As the applications get moved into virtual servers, it doesn't make sense to deploy iron in front of them, he added.

LineRate Point targets HTTP, HTTPS and Layer 4 load balancing, ignoring protocols like UDP (which would be mandatory for kit standing between the enterprise and the outside world) in favour of simplicity.

The virtualised load balancing also suits a world in which companies move toward automated orchestration, since the LineRate instance can be started and stopped under the same framework.

While the company wouldn't discuss the detail of pricing with Vulture South, Rafii said “typically … we're competing against free and open source solutions like the Apache Traffic Server.”

The volume licensing that goes with LineRate Point is designed to fit that strategy. Users select licensees on throughput (250 Mbps or 1 Gbps per instance) and per-instance (a load balancing deployment would count as two instances).

“There's a full REST API, and both the CLI and GUI are clients to that API. Everything you can do is exposed in REST to allow us to integrate to various orchestration systems”, Rafii said, citing environments like Puppet, Chef and OpenStack as examples.

LineRate supports multiple hypervisors and bare metal deployments, and is designed for a low per-instance footprint (a minimum of two virtual CPUs and 1 GB of memory; for a network boot, it can run diskless, or on a minimum 3 GB of disk space for a virtual disk install). ®

More about

TIP US OFF

Send us news


Other stories you might like