Without a high performance network, your data centre Is nothing

Connectivity is the critical concern for modern data centres, yet many co-location providers and data centre managers overlook the importance of high performance infrastructure. You can have access to all the storage capacity, computing power and intelligent environmental controls in the world, but without a low-latency, performance-driven connection between customer locations, redundancy sites and carrier partners, the quality and value of the service you provide is limited.

A user’s cloud experience is only as good as the weakest link, and this is often the network. It is crucial to bring network performance in line with the rest of the service’s performance.

Ironically there is a huge amount of telecommunications choice available to data centres and cloud operators, but many choose networks that lack the flexibility, latency, reliability and security necessary for global competitiveness. This is because operators do not sufficiently check the network they are implementing is actually right for them. This is even more relevant for providers servicing regulation-heavy sectors that have specific performance requirements – i.e. financial, retail, media and insurance.

Like resilient power and cooling, connectivity is essential to providing consistent services to end users and without it, data centres are effectively useless. However, what is an ‘appropriate connection’ in the context of the cloud. Key questions include:
· What characteristics are really needed per application – the cloud and data centres are application driven so how can the right network settings, in particular latency and uptime, be implemented cost effectively?
· What network standard offers the necessary functionality, value and global support?
· What new developments exist in the industry, and do these threaten existing investments?
· What complications can arise when undertaking a cloud-connectivity roadmap?

Realistically there is never a future proof point to invest in your network, but now is definitely better than most, especially as customer expectations around cloud connectivity reach a point where downtime is totally unacceptable, and mobile/data driven services herald in a new era of data centre utilisation.

The Four Pillars of The Network
Whatever network standard you choose, there are four considerations that must be addressed – bandwidth, availability, latency and long-term value. If you underestimate the importance of any of the above, catastrophic effects could occur. Reduced uptime and then a loss of customer loyalty, which filters through to your direct revenue stream, are just a few.

However, every data centre has different needs – some may need less bandwidth than others, some might only service a single geographical region and have a higher latency profile compared to, for example, a co-location provider that serves time-sensitive trading organisations.

The important thing is to invest in a network that is right for your facility and its growth expectations. Scalability and general availability are important for every data centre and should never be compromised on, but within such a fragmented and competitive market, what is the right choice? The sheer range of choice creates challenges of its own, especially when every provider is seemingly offering the same benefits.

The best approach is to recruit the assistance of a vendor agnostic provider; a connectivity expert that has built its sales strategy around customer satisfaction, bespoke solutions and consolidated support structures, as well as a focus on network-related return on investment and long-term value. By diversifying its portfolio across a range of global carriers, and providers, local access partners and co-location partners, a neutral provider will ensure your infrastructure objectives are at the heart of its technology decision, rather than its own profit margins.

Customer Expectations
For the customers connecting to your facility, you need to ensure you meet their requirements as much as your own. This is where the aforementioned expert can assist. They will understand the end user’s needs, whether it be network cost control, regulatory / compliance adherence, load balancing, globalisation, business continuity and redundancy, virtualisation, or general cloud utilisation for business / IT transformation.

That is not to suggest your internal operations team will lack the knowledge / best practice understanding to service your customer base, but often connectivity requires a unique approach.

This also applies to enterprises, especially if a company is setting up its own data centre rather than co-locating. In this instance cloud connectivity is critical, otherwise the investment will struggle to provide its returns.

No matter what phase of cloud maturity a business is at, every organisation has a specific network requirement. This applies to each specific cloud application as well. Thorough analysis is always needed. For example, cloud video is more intensive than a remotely hosted customer relationship management (CRM) platform.
Latency profiling, the volume of data being transmitted, and interpreting the effects the application will have on other critical applications are all extremely important. For the cloud provider, possessing a level of real-time visibility to provide this information to customers, as well as the actual network capacity to deliver a reliable service, will win business and retain existing customers.

Security is another concern, especially if customers operate within an audit-heavy sector. For instance, when financial organisations look to the cloud they are expected to ensure synchronous replication – regulators can enact large financial penalties if this is not in place. You need to identify these requirements and ensure your network is not lacking in any department.

Going Global
Another pressing challenge is the facilitation of IT globalisation. Enterprises are looking to the cloud for cost effectiveness and unless compliance demands it, often do not concern themselves with where their cloud solutions are coming from. They only want the business benefits. Therefore you need a network that can connect to customers across a range of locations, including regions that might lack the infrastructure of traditional IT hubs.

For enterprises looking to keep their data centre operations in-house, deploying in several locations for the sake of redundancy, as well as competitive advantage, poses a different approach. The connectivity needed in this instance is determined by a combination of risk management, cost, application type, scalability and latency.
Take storage replication as an example – it is fairly straightforward compared to process control or enterprise resource planning (ERP) systems. In contrast, non-critical applications need even less than a synchronous replication of data - a transatlantic connection of 60 – 80ms is a common figure. Every application is unique, so once the overall network deployment has been completed, specific application management should follow, bearing in mind the below:
· Bandwidth – a one-size-fits-all approach is needlessly expensive; implement bandwidth profiles for each application taking into consideration peak loads and growth that might occur unexpectedly
· Availability – avoid signing service level agreements based solely on availability administration; look for network design that results in network availability in line with that specific cloud service - services should be aligned to meet actual uptime levels
· Visibility – if a cloud provider / data centre does not understand what is occurring within their network, how can they guarantee a consistent end-to-end service? Ensure you understand the exact network topology providing the link to the cloud, whether fibre, Ethernet, IP or other standards
· Neutrality – objectivity means the best solution is picked for your specific data centre, cloud and application needs

Clear Skies
Application-to-application configuration is vital to ensuring the cloud yields its full potential. The cloud / data centre is only as strong as its weakest link, which often emerges as the connection within the facility. Searching for SLAs with facility-wide high performance connectivity will result in a stable, business enhancing experience, rather than one that fails on its promises. This should be a primary concern for enterprises looking to cloud enablement and data centre overhaul.
For data centre operators, positioning the network at the forefront of your growth strategy will ensure an investment roadmap / service objectives that result in a more satisfied customer base, greater overall reliability and a facility that actually delivers on its full potential. Holding back on the network will only cause complications in the long run.