WAN optimizer technologies to counteract data center network hogs

Off-site data hosting, rich media for business apps, and additional users will bog down your WAN. WAN optimizer and accelerator technologies can help.

Off-site data and remote users will tax your WAN, but there are several WAN optimization options to combat network overload.

It's hard to maintain fast and reliable wide area network (WAN) connectivity in the face of increasing data center workloads, user bases and real-time traffic. This acute demand for WAN services, combined with more geographic separation between physical data centers and/or service providers, creates bandwidth and latency problems. To address WAN bottlenecks, you can add costly WAN connectivity to the data center, curb network services to end users or deploy WAN optimizer technologies.

WAN bottlenecks and implications

WAN bottlenecks can delay every workload or end user to some degree. The typical culprits include data-rich files, communication between data centers, and remote users.

Large files such as high-resolution images, audio and high-definition video must be cached quickly. Even richly formatted text, PowerPoint and other business file types can contain enormous amounts of data.

Real-time voice or communication data -- Voice over IP (VoIP), Skype -- routinely travel over the WAN. Although highly compressed, these data types must transmit with higher priority than other data types.

Businesses regularly engage more than one corporate data center, sometimes in concert with outsourcing partners and public cloud providers. These satellite sites communicate with the primary data center and each other. All of the data transmitted for workload access, data synchronization, backup, disaster recovery (DR) storage and other tasks easily drives up WAN traffic and creates bottlenecks. Data centers can even experience problems synchronizing critical data, which will hurt hot-site DR performance and cause data loss.

As a company's user base grows, each additional remote employee, partner, client or customer accessing the corporate data center will add WAN traffic, especially as they use data-intensive and time-sensitive media.

WAN optimizer options

WAN accelerator technologies improve WAN utilization and ease bottlenecks. You can reduce WAN traffic with technologies that address specific attributes of WAN bandwidth, latency and traffic.

WAN bandwidth. Data deduplication and other data compression technologies convey data in fewer packets. Data deduplication -- particularly popular for storage tasks -- is gaining acceptance in networking. Deduplication replaces byte-level redundancies with small pointers that reference a single instance of the original data segment.

With local WAN cache, data transmitted over the WAN is cached at the receiving end, such as a branch office or remote data center. When the transmitting point receives a new request for the same data, it directs the request to the exact cache location. Since there is no need to re-send data that is already present in a local cache, WAN bandwidth use is reduced.

WAN latency mitigation. WAN latency occurs when data travels over large geographical distances between sending and receiving points and when communication protocols take time to acknowledge packets and packets are lost due to network congestion.

WAN cache addresses latency because it's faster to access cached data than wait for new WAN transmissions. Other techniques include using jumbo packets to transmit more data per packet over the network -- up to 9,000 bytes per packet -- and enabling selective acknowledgements, which allow the receiving end to skip acknowledgements for noncritical data packets.

Instead of resending lost or dropped packets, networks can be set to include loss-recovery packets preemptively after every few regular packets. The receiving end then reconstructs recent packets. This cuts down on the handshaking to request, resend and acknowledge re-transmitted packets in error-prone or congested WAN connections.

WAN traffic management. You can also optimize WAN by organizing and scheduling network content. For example, traffic shaping allows certain application data to take precedence over other data or prevents certain protocols or applications from flooding the WAN. You can ensure that business calls using VoIP get bandwidth ahead of corporate videos or press material downloads. Some traffic-shaping options provide network controls down to the user and application.

There are WAN optimizers that limit the number of connections and data rates, although you risk angering those users who are refused connections. Rate limits ensure that applications or users only use a set maximum amount of the available bandwidth.

Your choices in WAN optimization vendors

WAN optimization appliances and software typically require matching deployments at the sending and receiving ends. WAN optimization appliances include Dell SonicWall WAN Acceleration Appliance, Riverbed's Steelhead, Cisco's Wide Area Application Services (WAAS), F5's BIG-IP WAN Optimization Module and SilverPeak's NX Physical Appliances.

Many of these vendors also offer WAN optimization via software, such as Riverbed's Virtual Steelhead, Cisco's WAAS software and SilverPeak's VX Virtual Software.

Organizations can also explore open source WAN optimization offerings, including OpenNOP for Linux, WANProxy for data deduplication and TrafficSqueezer for Linux.

The features and functionality of each WAN optimizer vary, so consider the most beneficial feature sets for your network and put in the time to test and run proof-of-principle projects before deploying products into production.

This was first published in September 2013

Dig deeper on Data center network cabling

Pro+

Features

Enjoy the benefits of Pro+ membership, learn more and join.

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

-ADS BY GOOGLE

SearchWindowsServer

SearchEnterpriseLinux

SearchServerVirtualization

SearchCloudComputing

Close