16
SOFTWARE DEFINED NETWORKING
or more likely being reduced, so to have so much invested in
technology that is rarely used seems extravagant. Moreover,
most hardware or virtual appliances do not offer central
management or automation capabilities.
Indeed, the belief that specialised hardware appliances
were necessary was not helped when appliance vendors
tried to position their virtual appliances (which are deployed
on virtual machines instead of custom hardware) as the
solutions that enterprises should consider when they were
looking for a software-based alternative. Unfortunately
for the organisations that implemented these solutions,
these virtual appliances were not sufficient to handle their
performance needs and didn’t offer any other advantages.
Not only did they have the architectural limitations of
hardware appliances, they were actually too fragile to take
on and deal with sudden increases in traffic or dealing with
high volume encrypted traffic. On the other hand, many open
source software solutions, which are easy to download and try
for developers, are not feature-complete or fit for production
deployments. All of this has simply muddied the water and
further reinforced some of the market’s misconception about
the use of software to implement robust load balancing
architectures.
Powered by innovation
Advances in Intel architecture servers, with faster processors,
memory and both hardware and software improvements
in network cards, have enabled the construction of load
balancing solutions built on software-defined principles.
These software-defined architectures have allowed network
technology providers to look at the sector with a new set
of tools, which offer unprecedented levels of flexibility, cost
efficiencies and the robustness needed to deal with modern
application and network requirements.
These advances in application delivery architectures
empowered enterprises to rethink what’s possible with
load balancers. For years, innovation in the space had been
stymied due to legacy hardware vendors such as F5 Networks
and Citrix NetScaler that have done little to address the
needs of modern applications and cloud-native use cases.
Software-defined load balancers are agile and flexible enough
to scale up and down according to real-time traffic demands,
ensuring that organisations are only paying for what they use.
This flexibility is achieved without having any impact on the
performance of the network or any downtime.
This rethinking of the application delivery architecture
begins with the separating of the control or management
plane from the data plane. This allows central control over a
distributed pool of software load balancers that run on any
physical server, virtual machine, container or the public cloud.
Testing has proved that software-defined load balancers
can elastically scale applications from zero to one million
SSL transactions per second, using real-time analytics driven
decisions to automate the provisioning of virtual services
and the spinning up of servers. By scaling up load balancing
services during peak traffic times and scaling back down
when the traffic reduces, without any impact on the network,
organisations can now look to software-defined load
balancers as a viable alternative to hardware. And of course, it
comes at a fraction of the price as you’re only paying for what
you need and when you need it.
A pivotal industry moment
The level of encrypted traffic and seasonal events driving
traffic spikes isn’t going to change. In fact, businesses are
becoming more and more application-centric, adopting IT
processes and software development practices that enable
them to continuously roll-out new applications or updates to
existing ones. Having a solution with the elasticity to scale up
and down easily, combined with the robustness that ensures
the network isn’t going to fall down during periods of intense
activity, gives network administrators peace of mind and a
key tool to deliver better services to their internal customers.
The elastic scale afforded by the new architecture also
provides an unexpected advantage when dealing with
adverse events such as DDoS attacks. Auto scaling load
balancers based on traffic thresholds, means that networks
can manage the huge traffic surges associated with
such attacks until the threat has been mitigated. This is
achieved by scaling beyond the traditional data centre, and
automatically taking the extra traffic into private or public
clouds. The software-defined architecture for L4-L7 services
has now come of age and offers a robust, viable alternative to
the cost prohibitive hardware solution. n
www.networkseuropemagazine.com