load balancers &
application performance
The volume
of online
transactions
has continued
to grow
exponentially
over the past
few years across
an increasing
number of web
enabled devices.
balancers that are correctly sized
in front of individual applications
or tenants.
This distributes the traffic burden
across multiple load balancers
and provides immediate isolation
of services enabling organisations
to reduce downtime for all shared
tenants due to the maintenance
window for one tenant. It also
reduces costs since it eliminates
unnecessary overprovisioning
of hardware. Additionally, with
centralised control, companies
can automate the scaling out
of load balancing services by
dynamically invoking additional
software load balancers where
needed. This boosts application
availability and performance.
Webscale performance
and new encryption
algorithms
Distributing load
balancing services to
achieve cost efficiency
Appliance based load balancers
which have been a staple of the
networking industry are proving
to be both an expensive and a
limiting choice for enterprises.
The continuous heavy investment
in bigger and better performing
load balancers is a real issue for
companies since these appliances
come at a huge cost and are
typically overprovisioned to
handle unexpected traffic spikes.
Since they rely on proprietary
hardware, these solutions are
not elastic and do not enable
automation of services. The ability
to mitigate such wasteful spend
could prove a huge positive. New
strategies for per-application load
balancing provides the answer.
Instead of deploying a single
load balancer to service several
applications, some organisations
are implementing centrally
orchestrated software load
‘Appliance
based load
balancers
are proving
to be both
an expensive
and a
limiting
choice for
enterprises.’
Advances in the performance
of standard Intel x86 servers
have enabled large cloud service
providers and web giants such
as Google and Amazon to deliver
reliable, high performance
application services. These
webscale strategies rely on low
cost x86 servers, distributed
architectures, software and
API driven automation, and
rich visibility and analytics.
Similarly, software load balancers
implemented on standard Intel
x86 servers are now becoming
a mainstream approach in
enterprises thanks to their price/
performance ratio. Software
load balancers imple mented on
standard Intel x86 servers can now
terminate more than 2,500 SSL
transactions per second (TPS) at
the lower end, to nearly 75,000 SSL
TPS on a 36-core Intel x86 server.
Compared to hardware based load
balancing, that means software
defined solutions can sometimes
incur just a tenth of the cost for
similar scale.
Whilst the RSA based public-
key cryptography remains
generally popular, it has been the
increasing adoption of Elliptic
Curve Cryptography (ECC) that
requires smaller key sizes to provide
equivalent security that has proved
a benefit for many. Smaller key sizes
are less computationally intensive,
and result in better performance of
servers handling SSL connections
and improve the overall number of
SSL transactions per second (TPS).
Using load balancers
for application
performance analytics
The load balancer occupies
perhaps the most privileged
position in a network (in line with
applications), but it is continually
underleveraged as a source of real
time application performance
information. Load balancers
could play a much more expanded
role in application networking
services with architectures based
on software defined principles. By
using a distributed data plane of
high performance software load
balancers (that run on standard
x86 hardware) orchestrated by a
central controller the system can
collect and analyse millions of
continuous application data points
in real time. This enables the load
balancer to effectively provide a
rich amount of application insights
creating a ‘Tivo-like’ experience for
reviewing application analytics.
This means that an essential
component of your network also
becomes a critical application
analysis tool. It gives network
administrators a unique network
based perspective of application
performance, security posture and
end user experience. By zeroing
in on application performance or
connectivity issues, application
developers and networking teams
can work together to expedite
troubleshooting.
May 2017 | 35