Exploring the balance between bandwidth management and cloud computing

Bob Deutsche joined Intel in 2004 with more than 25 years of business and it experience as head of data center business, software development director and CIO. You can find more information about him in the Intel Server Lab.

Like many others around the world, you may have visited Los Angeles, especially this time of year. Whether your aim is for business or pure travel, you are likely to choose what we call the No. No. 405 Freeway (also known as the San Diego Expressway). The final renovation of Highway No. 405 was completed in 1969 and is currently one of the five most congested highways in the United States, and its design has been badly adapted to today's traffic.

This analogy may not be appropriate, but we can still use the case of Route No. 405 to begin our discussion: bandwidth and data transmission may not always be as cheap as it is today.

I've explored cloud policy in my column, and bandwidth management is what you (and your cloud provider) must know about the regulatory strategies and standards for cloud computing. Given that every form of regulation has its own way, the best advice I can offer you is to balance your goals with your regulatory policies. There is absolutely no one-size-fits-all template.

Broadband Spectrum considerations

First, we must recognize that broadband spectrum is a shared resource.

As a consumer, I certainly want all family members at home to be able to use any number of devices while downloading our favorite movie or http://www.aliyun.com/zixun/aggregation/17382.html "> TV series, and turn on instant audio and video chat, keep track of our stock portfolio dynamic trends, and play online games. I also want the delay to be less than 100 milliseconds, and the cost aspect should at least make me feel "fair". And with the choice of different suppliers, also let our experience gradually grow.

As a business, I understand that cloud computing data centers need broadband connections to any number or type of devices, including those 0/thin clients that can only receive data. In addition, these systems must provide the appropriate service and reliability to meet specific business needs. For example, health care and command and control systems must ensure their reliability and low latency requirements. Given the need for quality of service and reliability, I may need to pay vendors to disclose meaningful data, such as terms of service, pricing, packet loss, latency, service availability, and SLA based on response time.

Evolving revenue Models

With this in mind, it may be that the revenue model for broadband services will soon change (a big change due to geographical differences). I saw this chart earlier this year at Telco2.0 's press conference:

The image shows that the 2016 mobile traffic will be increased by about 15 times times than today, mainly video traffic. The video traffic forecasts I saw recently at the Gartner Data Center conference in December 2011 suggest that this figure may be conservative.

Historically, voice services have become a cash cow for telecoms companies. But the above figure also shows that the growth of this flow is expected to be flat. The above figure emphasizes that today's broadband fees are based on time estimates, not on network usage.

A large number of data streams on the network

Participants in the meeting discussed the slide with an implicit consensus that this pattern could change, given the amount of data (wired and wireless) that is being transmitted over the network.

There is some evidence that these changes have occurred at least in the United States. In the May 2011, a large U.S. operator began imposing monthly data restrictions on its fixed broadband subscribers. The move was very unpopular. Again, perhaps in security terms, tiered pricing for how much data or data is transmitted at what speed, or the blending patterns of both will evolve over the next few years. Can I be absolutely sure? No. But what if this phenomenon happens? How can I mitigate the company's risks? Especially how can I prove the return on investment in cloud computing?

Another topic that is important in the United States (mainly because, as far as I know, has no jurisdiction over wireless service providers), involves network neutrality. Network neutrality is based on capping the data. Discussions on this issue are best left to lawyers, and you and your cloud service provider need to be at least clear about what happens within the framework.

My advice, however, is to carefully anchor any business activity that expands your data center cloud and extend to your choice of how to design your cloud-based application portfolio and intelligent end user terminals. At least know if this will affect your cost and success.

Finally, to help you better smooth out the ins and outs, I've sorted out a complete list of the basic truths of the Enterprise cloud Computing strategy:

Large-scale transformation to cloud computing, including critical business systems, is a long-term process that takes 8-10 of years.

The L cloud is a top-down structural framework that combines the development strategy of the solution.

• The robustness and adaptability of your cloud ecosystem is only part of it.

L Service-oriented enterprise classification is not optional.

L cloud computing is a verb, not a noun.

• Technology-oriented business practices tend to circumvent government regulations, but legal and government policy standards will determine cloud success.

L Bandwidth and data transfer may not always be as cheap as today

(Responsible editor: Lu Guang)