By Malcolm Orekoya, Network & Security Specialist, Netutils

Views expressed in this post are original thoughts posted by Malcolm Orekoya, Network & Security Specialist, Netutils. These views are his own

For decades the availability of networks to their intended users has always been a priority. Data centres, from small to large, from enterprise to service provider networks, have always been built so that their infrastructure; routers, firewalls, switches and everything else in between, had layers of redundancy, resiliency and high availability.

As more and more applications get developed and more services move to a hosted or cloud infrastructure, we’ve seen over the years various guarantees of service availability. First it was 99.9% availability, then 99.99% and now we regularly hear of the five nines; 99.999% high availability, guaranteeing downtime of less than 5.26 minutes a year (if we keep going in that direction, nine nines availability has a ring to it I think).

So why all the nines, what does it all mean to your business, how can your business guarantee such high availability, how does this all connect to the cloud trend and what is the impact on your users? Well to your business, it obviously means you get less downtime in a year and as a result your critical services and applications are available more than they’re not, which hopefully means users productivity is maintained or increased.

One of the key considerations for enterprises today is whether to keep managing their infrastructure, services and applications in house (that is within their own managed data centres) or whether to move some or all of it to the cloud (that is a hosting provider) in an effort to keep up with the nines. There are arguments to support all options, but the trend is very much moving towards offloading parts or all of certain business critical infrastructure, software and services to third party cloud hosting providers and a couple of the reasons for this are:

Costs – Reducing IT costs (OPEX and/or CAPEX) is always top of the agenda for businesses and being able to do this while maintaining the level of availability of services and applications is paramount. With technology continually changing and the amount of applications continually multiplying, it is an increasingly difficult and costly task to keep the network infrastructure that underpins business activity highly available. Costs of hardware, software, maintenance, support and staff to manage the infrastructure (training costs included) all add up and usually these costs reoccur at the end of products or systems lifecycles, which is typically between three to five years.

By moving applications, services and infrastructure to the cloud, enterprises can take advantage of the robust infrastructure that these providers have in place, to provide multiple layers of resiliency, redundancy as well as compliance without having to incur the high costs associated with achieving this themselves. In many cases enterprises are able to reduce capital expenditure (CAPEX) for projects and introduce more predictable operating expenses (OPEX).

Finally, at least enterprises know that their availability SLA with cloud providers means they have someone else to point their finger at and in some cases get compensation should they experience any downtime. Although I would not base any decisions on which solutions to deploy based solely on this.

Flexibility to change – A lot of enterprises build in a degree of scalability into their systems from the beginning to allow for growth. But as technology, software and user behaviour changes so rapidly nowadays, enterprises can frequently find themselves inflexible to change or upgrades that could be beneficial to their business and can often find that the degree of scalability they had accounted for might not be enough. In addition things like product and system life cycles, which as I mentioned earlier is typically between three to five years, after which time the platform is considered out of date in comparison to the latest trends, can result in enterprises being forced to either spend a lot to get their platforms up to date or having to lag behind for a while.

In a cloud environment the speed at which new technology, systems, software and application updates can be made available to users is much quicker and one of the reasons for this is that many providers have a direct relationship or partnership with vendors that allows them to use and test various platforms well in advance of general release, thereby allowing them to leverage this vendor relationship by being able to quickly pass on new products, technology and services to their customers, while still maintaining the 99.999% availability of the system to the end users. The cost to the enterprise is normally a minimal increase in OPEX as opposed to large costs that can be incurred with rolling out platform updates or upgrades. In addition the flexibility offered by the cloud providers, means enterprises can scale up the resources that they utilise much quicker and easier than they typically can with an in house system. For example a cloud storage-as-a-service offering can in many cases allow enterprises to double the amount of storage being used by simply submitting a support ticket request to the cloud provider as a predefined cost.

To summarise, taking into consideration the ways in which we now work, with mobile users, increased collaboration, cloud applications and the as-a-service boom; in today’s enterprises the end user is king (arguably so) and business productivity can almost directly be linked to the availability of the enterprise resources that are used on a daily basis by users. Therefore the number of nines present in the availability percentage of enterprise systems will most definitely remain very important and the more nines there are then the less downtime the system is technically going to experience. But the balance of who can provide the most reliable availability figures (in house or in the cloud) will always be tipped by various enterprise specific factors that have to be weighed individually, there is no one size fits all.