Intelligent Data Centres Issue 05 | Page 20

DATA CENTRE PREDICTIONS approach, organisations are limited to the biggest box available at the time. And if the monster switch they bought isn’t big enough, they then have to rip it out and replace it, which translates into downtime. Instead of buying a large, high-capacity switch, you can combine multiple smaller switches to get the performance you could get from one large switch. This is called scaling out or scaling horizontally. The advantage of this is that you aren’t limited by the power of a single switch. As your needs grow, you can add more smaller switches and they’ll all work together to share the load. A happy consequence of scaling out is that you limit the size of your failure domains. If one small switch out of dozens fails, the impact is small. If it has redundant connections, the impact may be minimal – just a temporary drop in performance. As one might expect, cloud providers use the scaling out approach to achieve performance, reliability and agility at a massive scale. If an organisation is serious about Digital Transformation, its network has to use the same approach. The pursuit of fully automated configurations Today, there is unprecedented pressure on the data centre network to not only scale and remain performant, but also gain far greater agility and flexibility through extensive automation of the entire network life cycle, from provisioning and deployment to day- to-day management and upgrades. As the network grows and becomes more complex, manual configurations become increasingly time-consuming, difficult and risky. Automation is therefore a key piece of the scalability puzzle. To make the network fast and agile, your operating system needs to support open source automation tools like Ansible, Chef and Puppet. Your networking team needs the freedom to craft customised automation solutions that meet organisational objectives. Staying on top of the health of data centre networks Combined with the threat of network outages, the widespread adoption of microservices, containers and virtual machines has added a new layer of complexity in the data centre, resulting in a strain on traditional networks and the need to achieve operational simplicity. When an issue arises, organisations are forced to go hunting for the proverbial needle in the haystack, implementing a manual, box-to- box intervention because they don’t have a holistic view of all the activities putting demands on the network. In order to ensure the data centre network is behaving as intended, obtaining a holistic view of the network is critical. This translates into the ability to collect data, analyse it and visualise it in real time in order to obtain actionable insights and effectively manage the data centre network. CEOs are increasingly viewing infrastructure as a strategic resource to their business and having a direct view across the network enables teams to prevent outages achieving an even greater ROI from their data centre. The benefits of challenging the networking status quo Enterprises in all vertical markets and geographies are finding that change is inevitable in the era of Digital Transformation. This is especially true in the data centre, where architectures, infrastructure, people and processes must go hand-in-hand with increased IT and operational efficiencies, faster provisioning of resources, improved network management and troubleshooting. All of this needs to happen within the context of faster time to market and better alignment of IT teams with the business outcomes that matter most to the organisations that employ them. Fortunately, IT transformation initiatives provide enterprises with the opportunity to reassess network architectures and operational models in light of the need for a modernised approach to IT. In the sphere of networking, that means implementing practices that heavily emphasise automation, flexibility and scalability to deliver operational agility and efficiency gains, which would ultimately translate into increased revenue and overall organisational success. ◊ Unfortunately, the tight coupling of hardware and software limits organisations’ automation choices. A proprietary network operating system means either using proprietary automation software or, as this is a closed environment, using an API to stitch it all together. Again, the ‘one size fits all’ mentality prevents an organisation from achieving its Digital Transformation goals. 20 Issue 05 www.intelligentdatacentres.com