Subscribe / Unsubscribe Enewsletters | Login | Register

Pencil Banner

Avoid moving from a dumb pipe to a dumb cloud

Jim Morin, Product Line Director, Managed Services & Enterprise at Ciena | Sept. 14, 2012
Back in 1998 when Google Search started to replace AltaVista, could you imagine the company would be laying fiber in Kansas City in 2012 to offer a 100 times faster Internet service? Or that the New York Stock Exchange would evolve from a trading floor in its 1903 building on Broad Street, to offering virtual and physical colocated computing infrastructure services in its New Jersey data center?

The New Service Provider will put all the pieces together through acquisition or partnership. Critical elements to success include:

" Cloud operations software stack integrating the cloud operating system, security, operations, management and billing

" Software-driven service automation, workload orchestration and performance-on-demand

" Inter-data center cloud on/off ramps and networking infrastructure, with hypervisor network virtualization and distributed control plane

" Network operations and management

" Information management tools

" Data center servers, storage, networking

" Data center infrastructure and facilities in multiple locations

This type of provider is only beginning to emerge as cloud applications evolve from the first steps of moving applications to the cloud under software as a service (SaaS), to using the cloud for infrastructure as a service (IaaS). Connecting to many SaaS applications can be accomplished with no more than a standard Internet service delivered by an ISP, or through a companywide shared Layer 3 VPN network.

But as organizations begin to use the cloud for IaaS, the network can potentially break down because standard Internet connections are not built to handle the substantial amounts of information in storage nor the virtual machines now being transferred between data centers. The New Service Provider will need to deploy Carrier Ethernet and packet optical systems to provide the scalability and low latency required by IaaS workloads. And network hypervisor software are essential to virtualize network capacity, which will provide more efficient multi-tenant provisioning, automation and workload orchestration.

The New Service Provider can then leverage the network to increase revenue potential by building new, "cloud-born" applications that operate seamlessly whether physically located in an enterprise data center or in a cloud provider data center.

Today the cloud offers an "unlimited" processing platform with better operating characteristics for cost per CPU cycle, output formatting, input/output and storage. Server virtualization now allows logical units of workload to be orchestrated back and forth in real time both within the data center and also to the cloud. This gives a developer the choice to swing a complicated or CPU-intensive process out to a cloud provider like Amazon for more efficient processing on Amazon's "infinite" amount of servers, then bring the results back in-house. Some of Amazon's customers use this capability for CPU-intensive graphics and video rendering tasks.

Other examples of "cloud born" applications will emerge as developers take advantage of the new cloud-based tools and advanced networking. In much the same way that the cellphone maturation to smartphones drove the development of fourth-generation wireless networks, these new cloud applications will drive networking beyond standard Internet connections to support the delivery of new cloud-based applications.

The pay-off

The New Service Provider will be able to capitalize on the next stages of the cloud era enabled by software driven networking. The next-generation applications will take advantage of tight integration with the cloud platform stack, new tools for security and management integrated through APIs, and tight network integration of on-demand bandwidth, automated deployment and service-level orchestration. Optimization of cloud operations over cloud backbone networks will provide the resiliency and performance management required of next generation applications.


Previous Page  1  2  3  Next Page 

Sign up for Computerworld eNewsletters.