IT comes naturally
0203 239 2476

Cisco UCS

Since entering the server market in 2009 Cisco have seen a rapid level of growth with their UCS product range going from zero market share to now being the number 1 blade server vendor in the US by market share.

To understand how this rapid growth came about it’s important to understand the environment that UCS was designed in and how Cisco’s approach differed from the competition and to also understand the design innovations Cisco introduced with UCS.
When you look at the traditional server vendors each have been in the market a long time and during that time the use of IT and technology along with design concepts have evolved. We’ve seen the shift from mainframe to client/server and the introduction of virtualisation. We’ve also seen an increase in compute power and memory capacity.

With each of these shifts traditional server vendors have been forced to evolve their technologies and architectures to try and fit the changes in IT demands in the most cost effective and profitable way. But is this the most effective way to solve a problem?

This is why when Cisco entered the market they were able to take a fresh approach to the server market. They were able to start with a blank slate. Looking at the issues that affect modern day IT departments and design a product to specifically solve them.

Some of the elements Cisco introduced with UCS were:

Unified Fabric

In legacy blade architectures you had three separate networks, LAN, SAN and management this meant you had 3 or more adapters in each blade for connectivity into three separate switches which led to a spiders web of cabling and management headaches as well as inefficient cooling and higher power costs.

In the modern data centre the efficient use of the three elements of space, power and cooling are essential. With limits on physical space, the high cost and limits on power drawer and the need for “green IT” there is a requirement to make your blade infrastructure as efficient as possible.

Cisco’s answer to this was to converge the three networks onto one physical switch (2 in a cluster for resilience) and to introduce converged network adapter in the servers allowing for a massive reduction in cabling which lead to more effective airflow and cooling and by making a reduction in physical switches there was also a massive drop in power drawer.

Unified Management

When looking at legacy blade architectures due to the number of independent devices and the way the products had evolved you were left with multiple tools to manage your blade environment. This led to an unneeded level of complexity. Cisco approached this by creating a single management tool (UCS Manager) that could manage your entire blade environment from a single GUI interface simplifying the management of your blade environment.

Unified Compute or stateless computing

With Unified Compute cisco took a concept more associated to virtualisation and introduced it to the physical devices. The blade servers in UCS are essentially bare metal or dumb terminal until they have a service profile associated to them. By creating a service profile you were able to create a personality of a server independent of the hardware and associate it a physical blade giving you the ability to migrate between devices.

The easiest way to think of this is similar to a mobile phone. Your physical mobile phone holds no intelligence it’s just a pool of resources and power. The intelligence and personality of your mobile, ie phone number, network, messages etc, all sit on the sim card. You can move this SIM card to any other compatible mobile phone and turn it on and still have the same personality, number and so on.

This is the same with UCS the blades being the physical mobile handset and the service profile being the SIM card.

Some of the other key areas Cisco developed while designing UCS was the ability to incorporate and manage both blade and rack servers under the same management domain and also the ability to expand your blade environment to a total of 20 chassis working off a single pair of converged switches (Fabric Interconnects).

As you can see there are multiple benefits and innovations within UCS with many more I haven’t mentioned. The key point is UCS has been developed as an innovative blade architecture to fix modern day IT challenges with a simplistic architecture.

by Martin Hegarty | Architectural Lead, Comstor

Share this story:

Comments are closed.