The traditional three tier switching model of Access, Distribution and Core was conceived by Cisco in the days where data thus traffic flowed vertically to a single point, gathering speed on the way.  100Mbp/s access to GbE uplinks aggregated on 10GbE switches at the core, these in turn fed big fat application specific hosts with GbE network adaptors pushing and pulling very similar data types – mostly all fat clients.

The world is now flat….. Well, its getting flatter by the day.  Applications will no longer live in private data centres; Enterprises will use a mix of SaaS and Private Cloud solutions to meet tomorrow’s demands.  Virtualisation will allow business to turn up and down IT resources to meet with demand.  This has a huge effect on your network infrastructure.

To add to your woes, your staff are using more and more Social Applications, not just Facebook and YouTube but Google Apps, Internet chat and Blogging are all seeing a huge increase within the Enterprise.

So what do I do?

The concern with the traditional model is latency – forcing packets to stop at every layer.  Enterprises should build networks with a distribution layer of 10 GbE switches that is flattened out, becoming the communication link between servers with as few hops as possible, thus killing network latency.

Leaf-spine topology for computing network architecture

Some describe this two-layer switching method as leaf-spine switching topology or, similarly, a fat-tree switching topology.  In this scenario, servers are connected to leaf switches, which are then connected to a broad web of spine switches that provide interconnected bandwidth between leafs and spines.

This fabric of switches, which includes as many ports as possible, allows equal bandwidth access to every connection, enabling non-blocked movement data in an any-to-any server environment. Cloud leaf fabric controls the flow of traffic between servers, while the spine switching fabric moves traffic between nodes bi-directionally.

Very little is static in a cloud environment. Instances of servers and networks are provisioned at the drop of a hat.  To this end, the network architect must seek out a partner that understands the end-to-end solution from Application to User.  Varidion are a next generation service provider that builds and operates fully managed Infrastructure as a service that allows Enterprises to focus on the business elements of IT and not the technology end.

Remember – The network is the Computer!

Private Secure Cloud – Enterprises will see through the Fog and embrace the “Private Cloud” and host ICT within their network and not at the edge of their network.  Be quick!  Hosting facilities are filling fast and some very good deals can be had.

Application Specific Clouds – You buy vertical applications for specific business need, so why can’t we do this with Cloud?  You can.  Subscribing to different vendors for Voice, Email, Storage and Security is the best world strategy but can prove complex.  Incidentally, this is a key benefit of Varidions HYDRA platform – Choice!

Self Service – Outsourcing an application to a Cloud provider doesn’t mean you lose control, however some providers have better management platforms than others.  The ability to control your applications and services via consoles and control panels is big business and as SaaS providers try to differentiate themselves, this is where the gaps will appear.

Quality of Service – Many early adopters cite “Internet Weather” as a major complaint of SaaS; when users complain of slow response from applications and erratic performance the Network and IT Manager is helpless as the problem is after their network but before the SaaS provider.  Varidions HYDRA leads the way in providing QoS based interconnects between Customers and SaaS providers.

Application Monitoring – you’ve subscribed to SaaS, your applications and services are in the Cloud and users are complaining.  You’re blind!  Understanding where your providers are meeting or exceeding their SLAs is critical to companies that are to embrace SaaS.  Not only that, what is the actual application performance users are getting?

Cloud or Sheep?

It’s all Cloud, SaaS, Cloud, SaaS…

But to someone who has always started  a topology workshop by drawing a “Welsh Lover ” in the middle of the whiteboard when discussing ICT architecture with customers it’s not a surprise.  What’s is a surprise, is the way the “The Cloud” is being used by every provider, be they Network, Application or Systems Integrator, most who haven’t changed their product in years, but market their “new” Cloud or SaaS based service…..   That will be the Herd Behaviour!

Money saved by the cloud

is the best quote yet I have heard about why Enterprises should move to a Cloud infrastructure  The best bit, it was said by Andy Bechtolsheim the Chairman of Arista Networks and Co-Founder of Sun Microsystems…is he’s not alone, the Analysts agree, IDC predicts that spend on Cloud based services will be 6x that of traditional spend for the foreseeable future.

So why the hype?  Its simple, very simple, very simple indeed.  But that is it….  It’s all about Simplification.

Taking your applications to the network and not your network to the applications will drive complexity and cost out of any infrastructure.  And don’t get confused about Software as a Service, SaaS is the most abused term in IT right now and embracing a cloud model doesn’t need you to run headlong into Internet-based applications such as Google Apps and Saleforce.com, leading Cloud adopters are building their own Secure Clouds on Next Generation Networks.  Speak to Varidion about the Collision of the Clouds and how you can migrate your legacy into our Cloud Networking Platform - The Hydra.

Applications have changed and so have the rules; speak to Varidion about how our Hydra, the Cloud Networking Platform can deliver your applications quicker, cheaper and more securely.

Well not a server as a CIO knows of it today, not even a blade, the cutting edge of servers as they currently see it…

Our view at Varidion is the world of virtualisation is about to collide head-on with the world of networking as the applications of tomorrow will be run directly on what we know today as an Ethernet switch.  The reasons are simple; as the enterprise continues to deploy highly loaded virtualised servers running hotter and hotter it’s the network that will struggle to keep up with the server, so we have a couple of options:

Improve the network – The quickest way to improving your virtualized data centre as its likely both the topology and the type of switches installed are causing high latency with inefficient topologies.  In the vitualised world, three tier switching models don’t cut the mustard, a flat, low latency, switching infrastructure is the order of the day.  Unfortunately, even the “new way”  has a ceiling of 10Gbp/s or maybe 40Gbp/s in the near future!

More servers – By sharing the network and processing load across more servers a greatly improve performance will clearly be achieved, but so also will the cost.  Every server will need a network port, power and environmentals.  Not the ideal option!

However, there is a third way!   Today’s data centre Ethernet switches deliver significant processing power and Varidion believe if each logical Ethernet port could provide enough processing power to support a virtual appliance, the problem is solved.   An enterprise could literally have hundreds and hundreds of mini virtual servers that will admittedly support a reduced number of session per server but a greatly increased number of servers all sending and receiving at 1 Gbp/s.

Consolidation will occur but who will prevail? Watch this space!