By Eric Troyer (Part 2 in a 4-part series)
To understand how a new reference architecture can help MNOs meet the challenge of cost-effectively scaling their infrastructures, it’s useful to understand the limitations of the current infrastructure.
A Single IP Service Provider
Many MNOs have been strangled by outsourcing IP connectivity, especially single source (or “single-homed”) purchasing of IP services from one provider. In this “best effort” model, MNOs have no direct connection to their key traffic sources, such as Google, Facebook and Pandora. This means they can’t be assured of optimal performance. And once packets move from their infrastructure onto the provider’s network, they lose visibility and can’t differentiate traffic in order to prioritize the delivery of voice, video, messaging, and mission-critical applications.
With a single IP service provider, an MNO is limited to relatively few Internet access points, which may not map to population densities. This results in heavy traffic congestion and over-utilization of existing connections, which in turn creates a high-cost cycle of provisioning more access in these same inefficient locations.
A single IP service provider also limits the MNO’s ability to negotiate better prices and creates a single point of failure for mission-critical services hosted in cloud environments.
MNO-Run Applications and Services
MNOs often house their own applications, such as voicemail, messaging, and storage, in multiple data centers located far from population densities. This invariably results in greater latency, which can lead to poor customer experience. In addition, it’s extremely expensive to manage multiple data centers in multiple remote locations, and the procurement process for bringing new services online is long, complex and costly. Ultimately, the transport of these applications across long backbone links puts additional strain on available capacity.
A problem that particularly affects MNOs in North America is the need to acquire backhaul from multiple providers that have no standard, well-publicized locations. Anecdotal evidence indicates that in many metros and situations, it can take three to six months to track down the provider with the best lit building relative to the MNO’s network footprint, and then figure out the best way to connect the two networks. This expensive and time-consuming process is often further complicated and made even more expensive by the involvement of third parties in connecting the services.
In Part 3 of this series, we’ll look at a more efficient architecture for the mobile infrastructure based on multiple neutral data center hubs, or “mobility centers,” that house dynamic ecosystems of service sellers and buyers that span the entire mobile value chain, from network service providers to social networking, digital content, cloud and financial services companies, to enterprises.