Defying convention: why cloud service providers need to rethink their approach to visibility on the network

04 December 2013 | Pravin Mirchandani

Cover

Pravin Mirchandani

Blog Author | OneAccess; CMO


Cloud service providers, by definition, like to centralise everything. So, installing intelligent equipment at the very edge of the network goes against almost every ideological instinct they possess. But as the market for enterprise cloud services evolves and demand for bespoke cloud solutions increases, this instinct breaks down.

Cloud service providers, by definition, like to centralise everything. So, installing intelligent equipment at the very edge of the network goes against almost every ideological instinct they possess. But as the market for enterprise cloud services evolves and demand for bespoke cloud solutions increases, this instinct breaks down.

The ability to monitor and control both up and down stream network bandwidth flows is fundamental to the successful delivery of cloud services. Providers of enterprise-class cloud solutions need full network performance visibility if they are to anticipate and resolve problems before they impact availability and disrupt the customer’s service experience. But obtaining the level of visibility needed is becoming increasingly tough. Network virtualisation has long since disconnected the physical servers and transmission paths from the network models that providers sell on to customers, meaning that network performance monitoring solutions must now marry together what are effectively two different systems. What’s more, as the enterprise IT market really starts to take the cloud seriously, service providers are being asked to create bespoke solutions that blend public, private and hybrid cloud models all in a single service contract. In short, it’s all getting rather complicated. And when a distributed enterprise, like a major hotel chain or a brand of fuel service stations, requests a single managed service contract that will connect multiple sites all over the country, each with different connectivity requirements, then the business of keeping tabs on all corners of the network suddenly becomes a real technical challenge.

Finding a sustainable solution to this problem requires a break with convention. Network and application performance data drawn from the customer’s site, right at the edge of the network generates far greater visibility on the performance of the network and application infrastructure and, as a result, the overall quality of the customer’s experience. Fortunately, this can now be achieved via a multi-service access router installed on the customer’s premises. As a result, WAN optimisation technologies, for example, delivered as a managed service via the router, enable cloud service providers to both identify and differentiate between network and application performance issues.
This data arms providers with actionable intelligence, enabling them to respond quickly to resolve a network problem, or to defend themselves (and their SLAs) against wrongful performance accusations leveled at their network.

Tinkering with the network’s edge may be unfamiliar territory for cloud providers, but in truth, this is a pragmatic play, not a shift in ideology. Every connected customer has an access router already, so we’re just talking about upgrading an already-essential piece of kit.
As a concept, however, it may take some getting used to. The first to benefit will be the providers with the vision to spot the opportunity and the nerve to drive the change through.