OPINION
By Barry Silverman, Senior
Business Development Manager,
R&M UK
Barry Silverman shares his thoughts on how cloud
affects data centres, cabling and networks…
www.rdm.com
According to Cisco, cloud data centres will be processing
78% of workloads within the next two years, and 59% of
cloud workloads will be Software as a Service (SaaS). We’re
already seeing an uptake of hybrid cloud, with almost half
of all large enterprises expected to have adopted mixed
deployments by the end of next year. The rise of cloud
computing has a vast impact on data centres – it extends
resources over vast distances, changing resource delivery,
user interaction, application deployment and how data
centres interconnect.
Once upon a time, data centres might serve just one
company and each server would offer just one service.
Today, infrastructure is shared and a wide range of systems
need to be supported. The benefits are greater flexibility
and availability as well as lower cost. However, there are
several points that need to be considered today, in order to
build the infrastructure of tomorrow.
Convergence
‘Network convergence’ is on the rise. Integrated pools of
(virtualised) computing, storage and networking resources
are increasingly being shared across multiple applications
enabled by highly efficient, policy-driven processes, greater
flexibility and more economical operation. This offers new
ways of designing data centres’ underlying resources. It’s
vital to understand exactly how these work with the cloud,
and how they extend into remote networks. In the future,
vast converged infrastructure solutions may encompass the
data centre, as well as the cloud ecosystem to which they
are connected.
Software-defined storage
A vast amount of enterprise data is expected to move out of
on-site data centres and into off-premise clouds. Softwaredefined storage (SDS) solutions must, therefore, be able to
work seamlessly with ‘old school’ enterprise storage systems
as well as cloud service providers’ systems. Some type of
unified architecture is required to manage on-premise as
well as cloud storage, preferable from a single interface.
Virtualisation
Cloud service providers providing services via virtualised
data centres will have to be able to guarantee uninterrupted
connectivity – this requires rethinking of network and
cabling architectures – for example, in a traditional
environment, not every server is necessarily connected to the
storage network. In a virtualised environment, hundreds of
virtual networks or virtual routing and forwarding instances
could be carried over a single link. If that were to go down,
thousands of virtual machines and networks could go down.
Cooling and energy
The rise of cloud is resulting in a push for greater energy
51
efficiency. Regulators and other stakeholders are calling
for data centres to use energy in the most environmentally
sound way possible. Building data centres in cold areas to
save on cooling is just one example. Using more energyefficient hardware, advanced monitoring and convergence
all play a part in lowering PUE. Data centre designers need
to (re-) consider the effect cloud technologies have on
hardware, density, virtualisation and so on, and how that
relates to energy management and cooling.
Uptime and continuity
The changing role of the data centre brought on by cloud
has consequences for data centre design. Until quite
recently, most content was delivered locally within data
centres, but today we see far more inbound and outbound
traffic, as millions of people worldwide access data and
services 24/7. Millions of individuals, businesses, institutions
and governments depend strongly on data centre
availability. 100% uptime is a must. DCIM tools support
this by offering visibility, if necessary across multiple data
centres. Understanding how cloud-specific disaster events
come about and how these can be avoided is vital, as well
as using the wide range of automation, intelligence and
monitoring tools available today.
Thinking ahead
The cloud allows instant connection of computers, access
to servers, storage and offers flexible and 24/7/365 on the
go access, no matter where you happen to be. Whether
for business or social use, we have driven the requirement
for on-demand, high bandwidth applications and services
and we want it instantly. And the Internet in the sky duly
provides for us.
As users, we expect the infrastructure be effective, flexible
and reliable. That is our demand to the providers in terms of
the actual data centre facility. Providing outage prevention,
scalability and manageability is high on the list for operators
and managers. And being able to dynamically manage data
traffic while maintaining service continuity and of course
security, is essential.
Design and products th at reflect this approach but also
provide density while not jeopardising flexibility, control or
manageability, are key. It’s essential to provide a cabling
infrastructure that looks as far into the future as possible.
With developments happening at the speed of light, the
data centre designer can’t afford to let the physical cabling
infrastructure to be the limiting factor.
Data centres will increasingly need to provide more
than storage. Customers will demand risk management,
redundancy, connectivity and recovery services.
Infrastructure will have to be flexible, scalable and reliable
to accommodate the cloud-based applications of today and
those of tomorrow. n
www.networkseuropemagazine.com