Datacenter Design: Optimizing Networks for Evolving Traffic

  • Source: IDC
  • Written By:
  • Published:
  • (Originally Published On:) )
Data center network design architecture is changing to reflect growing IT trends in distributed applications and virtual computing. New computing models require a flatter network that is flexible enough to handle unpredictable traffic patterns. Find out what that means for your company in this interview with Cindy Borovick, Vice President of IDC’s Enterprise Communications and Datacenter Networks programs.

Featured Software Research:

The Need to Rethink the WAN

Technological advances have evolved at an exponential rate over recent decades—we’ve seen the advent of the Internet, cloud computing, virtualization, and several other technologies. But the wide area network (WAN) technologies that are currently in use in most branch offices have changed little since the 1990s.

Business environments today are experiencing unprecedented growth and change, and IT departments are losing visibility and control. Traditional WAN technologies are a mismatch for modern branch networks that need to be up and running quickly. The demand for adaptability, awareness of costs, and the need to run and monitor business-critical processes are indicating the necessity for a new WAN technology paradigm.

New WAN services based on the latest technologies have not been deployed by traditional service providers, primarily because they require an initial financial outlay which companies are not prepared to make. But they are poised to succeed multiprotocol label switching (MPLS), and are the most viable alternative for supporting application performance while dealing with increased network traffic and the connectivity demands of cloud computing and mobile access.

In this white paper, the characteristics of WAN technology and the limitations of traditional WAN technology, including complexity, are explored. This paper also discusses some new approaches, such as a hybrid WAN, which are better suited to supporting fast-changing business and technology environments.  Read More

Optimizing Bandwidth: Why Size Doesn’t Always Matter

Network managers have to balance the need for increased speed of response and the best possible end user performance, and the need to reduce costs.

Traffic on the network expands to fill the available bandwidth, resulting in degraded performance across the most congested links. Network managers have to be able to show that a link is experiencing congestion for a significant amount of time due to legitimate business usage before they can add additional capacity. But approaches such as peak or average utilization and traffic totals provide limited insight into and understanding of network congestion.

To optimize the use of existing capacity and make the business case for more bandwidth, network managers need to ask themselves a number of questions. In this white paper, a new approach to getting needed information about network congestion is discussed. Principles to consider for effective capacity planning are discussed, as well as the limitations to traditional approaches to managing network capacity, and a solution called burst utilization is given, to measure performance and save network managers time.

A burst-style view shows the links spending the most amount of time in a congested state and the time in that state. Using this information, the network manager can explore why the link is busy and whether additional bandwidth is needed, and identify any links where capacity could be reduced. By using existing bandwidth more effectively, they may be able to improve the end user experience.  Read More

Retail Retold: How the Internet of Things Is Evolving Omnichannel Retail

Most merchants and consumers have adjusted to the multi-device, multichannel environment known as omnichannel retailing. But there are real challenges that the omnichannel environment brings to the retail industry, even before considering how the Internet of Things (IoT) will amplify the complexity. P.J. Jakovljevic and Raluca Druta report on how the IoT is altering the retail world. Read More

You may also be interested in these related documents:

Green Server Design: Beyond Operational Energy to Sustainability

  • Source: HP
  • Written By:
  • Published:
“Green” server and datacenter design requires a focus on environmental sustainability. Prior studies have focused on operational energy consumption as a proxy for sustainability, but this metric only captures part of the environmental impact. Understanding the total impact requires examining the entire lifecycle of the system, beyond operational energy, to also include material use and manufacturing. See how one methodology can do that. Read More

NGN Control Plane Overload and its Management

Given the types of overload scenarios encountered within traditional networks, similar traffic principles requiring overload management will be required within (NGNs). What, then, are the characteristics of overload protection within real-time communications servers? And what types of overload management methods are available for optimizing server performance? Read More

Optimizing Bandwidth: Why Size Doesn’t Always Matter

Network managers have to balance the need for increased speed of response and the best possible end user performance, and the need to reduce costs.

Traffic on the network expands to fill the available bandwidth, resulting in degraded performance across the most congested links. Network managers have to be able to show that a link is experiencing congestion for a significant amount of time due to legitimate business usage before they can add additional capacity. But approaches such as peak or average utilization and traffic totals provide limited insight into and understanding of network congestion.

To optimize the use of existing capacity and make the business case for more bandwidth, network managers need to ask themselves a number of questions. In this white paper, a new approach to getting needed information about network congestion is discussed. Principles to consider for effective capacity planning are discussed, as well as the limitations to traditional approaches to managing network capacity, and a solution called burst utilization is given, to measure performance and save network managers time.

A burst-style view shows the links spending the most amount of time in a congested state and the time in that state. Using this information, the network manager can explore why the link is busy and whether additional bandwidth is needed, and identify any links where capacity could be reduced. By using existing bandwidth more effectively, they may be able to improve the end user experience.  Read More
 
comments powered by Disqus