As data center virtualization continues to grow, research from the Yankee Group finds moving to a fabric architecture as the new requirement in order to build networks that are scalable, simplified and able to meet all of the new challenges of the virtual data center.

Over the next couple weeks, we will look at the evolution of the data center and how the need for network fabrics has grown in importance. Using research by the Yankee Group in their article “Datacenter Solution is Dependent on a Network Fabric” by Zeus Kerravla, we’ll touch on the following topics:

• Evolution of the Data Center
• Limitations of Existing Data Centers
• Defining a Data Center Network Fabric
• Benefits of a Network Fabric
• What to look for in a Network Providers

Check back Monday, May 2nd for the first in this new data center series.

The final question IDC analyst Linda Borovick tackles in her “Datacenter Design: Optimizing Networks for Evolving Traffic” article is how does an IT organization balance the need for Layer-2 versus Layer-3 datacenter network design?

Ms. Borovick starts with the fundamentals:
• Evaluate the architecture of the current applications running
• Take a practical approach to creating a flatter network
• Ensure that the network can manage both physical devices and the virtual traffic moving between those devices

The challenge is that flat networks must have redundant connections. Layer-2 datacenter network design must have the attributes of Layer-3 in order to optimize available bandwidth without paying an additional performance tax.

IDC concludes that this new datacenter network design must be forward-thinking to not only include traffic patterns within the datacenter, but also accommodate the ability to securely handle communications between multiple datacenters – either private or public clouds.

IDC reports that the overall datacenter switching market is valued at approximately $4.8 billion with the ethernet aggregation tier between a fifth or fourth of that.

Although network managers are evaluating new datacenter network designs that optimize network traffic due to the need to create efficiencies, drive revenue growth and an increase focus on consolidation and virtual management, IDC analyst, Linda Borovick believes that the fundamental drivers for spending on networks have not changed.

Ms. Borovick points out the growth of distributed applications, voice and video, network-based business, network-attached devices, virtualization, and cloud computing will create demand for intelligent networks that support the notion of fabric computing.

On Wednesday, April 27th check out the last article in this datacenter series, “Striking a Balance between Layer-2 network design and Layer-3”.

Today’s datacenters are moving away from a hierarchical method to an environment in which systems are in a much more equal partnership writes IDC analyst, Linda Borovick in her article “Datacenter Design: Optimizing Networks for Evolving Traffic”. She further explains that because servers, storage resources, applications and so on are all equal, the datacenter network must be architected to treat them as equal in order to properly reflect the traffic flow.

Ms. Borovick highlights that next-generation datacenter designs must now consider the following:

• Server-to-Server and server-to-storage traffic
• Implementation of tiered-applications. In other words, the Web server, the application server and the database server are each running on virtual machines or separate physical servers.
• Increased complexity between individual virtual machines

The next post in this series “Market Spend on Ethernet Switches in the Aggregation Tier” will post on Monday, April 25th.

Linda Borovick an IDC analyst points out that there is an increasing reliance on the physical or virtual servers to act in a single fabric instead of silos of resources. So instead of network traffic flowing from a “single server to the end user” model, you increasingly have datacenter traffic that flows between multiple servers and data stores before flowing out to the end user making network traffic patterns more difficult to predict.

As data center virtualization and software-as-a-service, and ultimately cloud computing, continue to grow, end users are going to not only need to access applications and data from the corporate datacenter, but also from one or more service providers or business partners. Companies will see increases in bandwidth and capacity requirements.

The next post in this series “The Effect of Any-to-Any Server and Storage Capability on Datacenter Architecture” will post on Friday, April 22nd.

As IDC analyst Lucinda Borovick describes in her article “Datacenter Design: Optimizing Networks for Evolving Traffic”, ten years ago datacenter applications had a monolithic structure with a particular goal, a server with client computers attached for particular tasks which essentially created individual silos of computing resources.

Today, datacenter applications pull bits of information from individual locations and compile them to meet a specific need which has led to a rise in datacenter virtualization. This change requires data to exist through a centralized view to increase efficiency and cost-effectiveness and also to make data disaster recovery easier to plan.

Check back on Wednesday April 20th for the next post in this datacenter design series – “How is Network Traffic in the Datacenter Evolving”.

Evolving Solutions understands the changing needs and new challenges for datacenter management. A recent article from IDC, highlights one of these new challenges – optimizing networks for evolving traffic.

As IDC explains, datacenter design is fundamentally changing as new computing models require a flatter network that is flexible enough to handle the sometimes unpredictable traffic patterns caused by less hierarchical, more peer-determined computing in virtual and cloud environments.

Over the next couple weeks, we will be posting a Q&A with IDC analyst Lucinda Borovick which will touch on the following questions:

• How have application compute and storage architectures changed over the last decade?
• How is network traffic in the datacenter evolving
• How does the desire for any-to-any server and storage capability change the architecture of the network?
• How much does the market spend on Ethernet switches that sit in the aggregation tier of the network?
• How do you strike a balance between Layer-2 network design and Layer 3?

Check back on Monday, April 18th for the first part of this Datacenter Design series.

IDC points out that the recent increase in server virtualization to create a more dynamic and agile datacenter has led to an increase in complexity and costs associated with managing the network.

How does an organization keep total cost of ownership low? IDC Analyst, Matthew Eastwood, recommends automating network orchestration and freeing up human resources to focus on the right priorities in the data center which will lead to operating consistency and visibility across the server and network infrastructure. Data center networks built on these principles will manage capital and operational costs while providing flexibility, security and scalability necessary to tackle business challenges.

As an early adopter of  Server Virtualization, Evolving Solutions also understands the benefit of providing its customers open source virtualization.

As IDC Analyst, Matthew Eastwood, explains in his whitepaper titled “Managing Server Virtualization on Datacenter Networks”, with virtual environments becoming increasingly heterogeneous, an open platform leverages publicly-available APIs to communicate with virtual switches provided by a variety of hypervisor suppliers. Interoperability enables partners to build and deploy applications that are smart, simple and run on any physical server or operating system; whereas proprietary solutions built on a vendor’s own platform can reduce business agility and options.

Don’t miss the last post in this series on Server Virtualization on Wednesday, April 13th – “Impact on Total Cost of Ownership”.

The increase in server virtualization requires server and network administrators to automatically provision physical networks so they’re in sync with virtual networks in order to prevent policy-enforcement issues and virtual machine (VM) mobility failures.  In a recent whitepaper titled “Managing Server Virtualization on Datacenter Networks”, IDC Analyst, Matthew Eastwood, provided the following guidelines for an optimal approach:

• End-to-end physical and virtual network orchestration
• Automatic orchestration and elimination of errors caused by out-of-sync physical and virtual networks
• Unified network and security policies between virtual and physical networks
• Scalability and performance to manage large number of hosts

Check back on Monday, April 11th for the next post in this series – “Open Versus a Proprietary Solution”.