 | Issue: | North America I 2014 |
| Article no.: | 6 | |
| Topic: | Five key trends for network transformation in the ZettaByte era | |
| Author: | Nicholas Ilyadis | |
| Title: | VP & CTO Infrastructure & Networking Group | |
| Organisation: | Broadcom | |
| PDF size: | 194KB |
About author
Nicholas Ilyadis is VP and CTO of the Infrastructure & Networking Group (ING), in Broadcom Corporation. In his role as VP and CTO for ING, he leads the product strategy and cross portfolio initiatives for a broad portfolio of Ethernet chip products including network switches, high speed controllers, PHYs, enterprise WLAN, SerDes, silicon photonics, processors and security.
Prior to Broadcom, Ilyadis served as Vice President of Engineering for Enterprise Data Products at Nortel Networks and held various engineering positions at Digital Equipment Corporation and Itek Optical Systems.
He holds an MSEE from the University of New Hampshire and a BTEE from the Rochester Institute of Technology. Ilyadis is a senior member of the IEEE and contributes to both the IEEE Computer and Communications Societies.
Article abstract
As the ZettaByte era is approaching, five technology areas are developing to cope with it: SDN, NFV, Cloud, Big Data, and Green Data Centers. SDN enables network complexity to be handled by automated software based system; NFV hosts networked functions on virtualized machines, decoupling the functions from the hardware; Cloud serves both data and applications from a distributed networked environment, to any device, anywhere; Big Data techniques make it possible to perform real-time analytics on the large accumulated data, and Green Data Centers apply methods of minimizing the consumed power.
Full Article
Times are changing. Computer networking is no longer the patchwork construction it once was. Today, computer networks are the foundation for all modern communication. Businesses large and small, and individuals, are now all connected in one form or another to a modern network system. As a result, these systems and the data centers used to house them are being scaled up to accommodate the load. They are getting bigger, more pervasive and increasingly, more complex. The amount of data carried by networks has exploded, heralding in what many now call the Zettabyte era (1 ZB= 1021 bytes of digital information storage capacity). According to The Cisco Global Cloud Index, by the end of 2017, the annual global data center IP traffic will reach 7.7 ZBs. That is where the problem begins. How do companies and service providers manage the growing complexity and ever increasing need for bandwidth, while still enabling easy network configurability? A number of emerging trends now offer a way to transform the IT infrastructure, while also making it more manageable and configurable.
Software-Defined Networking (SDN)
Consumers want increased access to high-bandwidth, multimedia rich applications and content, and much more of it. This presents a host of problems for today’s networks and data centers—problems that are getting more complex with every passing day. SDN provides an answer to this dilemma. It manages this complexity via a network-wide software platform that enables centralized network coordination, control and programmability.
For IT professionals, turning the complex task of provisioning, optimizing and monitoring network traffic over to software presents a number of potential benefits, namely that it gives them a programmable and customizable interface for controlling and orchestrating the operation of a collection of devices at different levels of abstraction. This makes data center processes more agile and increases their performance. It also allows data centers to use their assets more effectively—all of which reduces cost and improves both efficiency and productivity.
While still in its infancy, there is little denying the impact of SDNs on the manual network configuration and management process. Being able to replace this effort with a software platform would certainly allow network administrators to roll out new services and functionality faster and with fewer errors, and more easily balance network loads. Such capabilities make SDN an enabler of other emerging trends, like network virtualization and cloud computing.
Network Function Virtualization (NFV)
Whereas SDN promises a way to manage and control increasingly complex networks, Network Functions Virtualization (NFV) grants users the flexibility to relocate network functions from dedicated appliances to general purpose servers, making networks more scalable, agile and efficient. Rather than buying a hard asset, such as a router or box for a single purpose, service providers can now take the function associated with the box and instantiate it as a virtual machine on a server. By using standard IT virtualization technology to consolidate many network equipment types onto industry standard high-volume servers, switches and storage, NFV aims to transform the way network operators architect and operate networks and network services (Figure 1).
Figure 1. The vision of NFV (Courtesy ETSI NFG)
The vision of NFV is to consolidate many network equipment types onto industry-standard, high-volume servers, switches and storage. While still in the early stages of development—the first draft of the NFV architecture document isn’t expected until January 2015—the technology promises a host of benefits. For example, because the network functions will be implemented in software, they can be easily moved to, or instantiated in, various locations in the network without having to install new equipment. Also, network operators and service providers won’t need to deploy as many hard assets. Instead, inexpensive, high-volume server infrastructure would be deployed with virtual machines running on top. Additionally, use of virtualization would eliminate the dependency between a network function and its hardware, allowing physical hardware to be shared by multiple virtualized network functions.
Currently some proof of concepts are underway for NFV, however SDN must first be in place for the technology to really take hold. As the industry begins its march toward virtualization, there will likely be a blend of soft and hard assets within the network, with many of the hard assets eventually being replaced with virtual machines.
Cloud Computing
When it comes to getting power for an electric device these days, it’s as simple as plugging it into a power outlet. We don’t think about where the power comes from or even how it gets to us, we just know it’s there in the outlet waiting for us to use it. Now, with Cloud Computing, this same concept is being re-envisioned for computing. Many of us have a laptop for work that’s carried home each night. What if computing became much like a plug-in utility? That’s the promise of cloud computing, with specific applications and data files available when and where they are needed, without having to be attached to any specific device.
Cloud computing has a number of benefits for the scientific community, business and society. It provides scientists with easy access to thousands of servers and the processing power they need to follow scientific pursuits. That same computing power can be leveraged by companies to test their designs and discoveries. Cloud computing even improves collaboration and makes education available to the masses regardless of location.
Big Data
Everything we do these days leaves a digital trace and there’s no shortage of people looking to use and analyze this information. ‘Big Data’ is the term many use to describe the enormous amounts of ever-expanding data—both structured and unstructured—that we generate on a daily basis. Collecting and analyzing this data has significant implications for businesses looking to monitor buying patterns, extract specific business information or make sound strategic decisions. It can also help to significantly improve our ability to understand the world around us, whether that means monitoring trends in social networks, foiling an act of terrorism or even finding a cure for cancer.
Organizations have been capturing and analyzing data for some time, however, the rate at which it is generated today has grown by leaps and bounds and continues to increase. Big Data analytics now allow those organizations to analyze very large, complex forms of data by breaking the task into smaller ones that can run in parallel on tens, hundreds or even thousands of computers within the cloud. By doing so, companies are now uncovering hidden patterns, unknown correlations and other useful information.
Green data center
According to the New York Times, data centers can waste 90% or more of the electricity they pull off the grid. With energy costs spiraling out of control and demand for computing power continuing to increase, the “greening” of the data center, i.e. using different techniques to reduce the operational power requirements, has become increasingly important. One possible green technique is usage-based power versus constant power. In the past, networks were always on, even if no information was flowing. They were powered up at maximum just to maintain link integrity between devices. In the energy-efficient networking world, however, links are now being powered up and down based on their loading. If a piece of equipment is idle, it powers down portions of its circuitry, and when needed, it quickly powers up again.
Using techniques like this, green data centers can realize lower energy costs and overall operating costs. Lowering data center power use also results in a potential CO2 emissions savings of up to 3.5 million metric tons (based on ~5 TWh/year), as shown in Figure 2. This infographic presents statistics based on the full deployment of energy efficient Ethernet on a population of links equal to those in use today, assuming all 10/100 links become 1G and account for some migration of data center links to 10G speeds.
Figure 2. Green Data Centers
Conclusions
With bandwidth demands exploding and data centers scaling up to meet the demand, finding ways to make networks scalable, manageable and configurable have become absolutely essential. The development of technologies like SDN and NFV to address this problem signal a paradigm shift in networking design and development that will increasingly be defined by software. Cloud computing will also play a role, redefining the way people use and access networks and data, while Big Data analytics is working to put the data to good use. For their part, green technologies will reduce adverse impact of the rapidly escalating power consumption. In all, these trends promise to transform not only the future of networks and the data center, but every aspect of society as well.
