The last five years have seen such a monumental shift in the data centre. Data centres have come a long way. The shift from rows of boxes with a 1:1 service to server ratio now reach 100:1. We have seen investment in storage, the arms race for the best servers, storage, and who can deliver the highest density 100 gigabit connectivity.

VMware have made leaps and bounds, especially with Hyper V and Openstack nipping at their heals. Openstacks relentless pursuit in the cloud space has allowed some health competition. With the shrinking costs of raw compute power reaching all time lows the ability to build a DC has become more affordable than ever. Application virtualization and the abstraction of historically hardware only services such as firewall, SLBs, and WAN acceleration to name a few, have been the catalyst to where we are now.

Cloud based architectures are generally pitched and spun through the marketing machines and sound like everyone else. When you get down to the nuts and bolts there are some common themes. SDDC aims to standardise through homogenous infrastructure. An adaptive, holistic framework that is flexible to support changing workloads and has an innate ability to self program and adjust for demand. Automation is the crux of a scaled out DC. How can a team deploy thousands of machines in a reasonable time? Via automating the mundane and delivering the awesome, that’s how! Resiliency caps of these nuts and bolts. Now that a software architecture can compensate for failing hardware, technology can empower businesses to achieve a new level of resiliency without the cost.

Screen Shot 2013-07-22 at 9.33.16 PM

In networks such as these, VXLAN, automation and flow control play their part. Heck, there is even layer 1 reprogramming. If an application needs to be moved around the DC then software drives the change. Software based technologies alleviates the need for human interaction when it can easily move applications business functions to where resources are pooled and active. No more hardware firewalls you say? Well at least, not as many! Don’t throw your firewalls out yet, but do consider the ability to run them inside the hypervisor. The ability to isolate and contain and perform firewall functions in software allows compliance, isolation, and a whole plethora of new designs to solve modern challenges.

Tiered storage used to take time. The business case to provision a new LUN or an entire SAN would be a hard sell. The prior planning that went into carefully carved house of cards would crumble at the slightest odd requirement. Profile driven storage aids in reducing operations complexity and increased agility. Storage DRS allows aggregation of many data stores into a single pooled resource. Flexibility is key – it is reducing the OpEx all time! Wrap all this up in a nice fluffy management single pane of glass and you have your SDDC. Get your finger on the pulse and identify and justify CapEx for long-term OpEx reduction. Enable charge back and drive revenue from otherwise hard to justify components. Make that 196 ports of 100 gig really earn its street cred.

Whilst SDDC loses its virginity to the marketing teams around the world, why not come see what the real players are up to at the Software Defined Data Centre symposium hosted by Tech Field Day, SDNCentral, and the Packet Pushers.

If you want to come along then register through this link. It will get you 5 dollars off your ticket!

One thought on “Software Defined Data Centre

Leave a Reply

Your email address will not be published. Required fields are marked *