Resiliency, latency, and security on the core Internet are problem areas for mission-critical cloud-based business applications, as are the irregular last-mile access networks that exist in many countries.
According to Ovum’s network infrastructure analysts, although cloud computing is a well-defined concept it will require a strong physical network infrastructure to live up to its promise. These will need improved interconnectivity for resiliency, load balancing, new transmission capacity; and added intelligence from new optical and IP hardware.
“Public IP will not be sufficient for business-grade cloud computing and carrier Ethernet and IP/MPLS VPNs will play increasingly important roles. More efficient data centers, and more of them, distributed globally, will also help,” said Matt Walker, Ovum principal analyst.
Ovum’s recent discussions with operators and vendors indicates that hope runs high, that IaaS (infrastructure-as-a-service) applications will generate more and higher-margin revenue - not just traffic growth - to support building multi-terabit networks.
Expanding on this theme, Dana Cooperson, Ovum’s practice leader, commented that “Public cloud IaaS business models are “pay-for-use” or at least tiered, which, unlike "all you can eat" consumer broadband models, tie resources used more directly to revenues generated. Public and private cloud business and pricing models are new and in flux, so there's still ample time to get it right...or wrong.”