Making broadband business-ready

Tying together remote offices has long been the playground of expensive, dedicated circuits, and lately, MPLS services. However, the spread of business-class DSL and cable Internet services has led to higher-bandwidth links becoming available in more places.

The downside to that is that most business-class broadband connections can’t be relied upon to deliver consistent bandwidth, which can cause reduced service and WAN problems.

Sampath Ramaswami is the Senior Director of Strategic Development at Hughes Network systems. In this blog post, he details how Hughes is addressing this problem by constantly measuring the available bandwidth of a circuit, then adjusting dataflows accordingly. — Paul Venezia

Making broadband enterprise-ready with end-to-end quality of service

MPLS has for years remained the preferred choice of many IT managers to connect their branch sites together — despite the limitation of speeds capping out at 1.5Mbps (T1), few evolutionary improvements, and bandwidth costs more than 10 times that of broadband. Meanwhile, broadband infrastructure, driven by major investments in recent years, is affordable, robust, capable of delivering high capacity, and available almost everywhere, with major providers delivered an average of 97 percent of advertised speeds, even during peak hours.

A key reason why enterprises are willing to pay the MPLS premium over the possible alternative of a broadband private network is that broadband cannot provide end-to-end quality of service (QoS)/class of service(CoS). While a residential consumer might be willing to tolerate slow performance on a key application (such as online banking) due to the impact of lower-priority traffic (say, a streaming movie), such a trade-off is not acceptable in an enterprise environment.

In the past, QoS over broadband end to end has not been achievable. Key parameters like capacity, latency, and packet loss may change abruptly as the network experiences congestion, without providing any indication to premises devices. Furthermore, bulk traffic is handled exactly the same way as real-time traffic, since the broadband network doesn’t honor CoS markings set at the edge.

WAN optimization with a twist

The good news is that viable solutions are now available for the next-generation distributed enterprise WAN over affordable broadband– without paying the MPLS premium. Hughes ActiveQoS is a performance optimization technology that enables real-time QoS over broadband connections, capable of converging real-time voice, video, and data.

ActiveQoS is WAN optimization with a twist. Whereas typical WAN optimization delivers QoS control, application acceleration, and increased virtual bandwidth, the QoS is effective only over leased access circuits such as a T1, with fixed deterministic capacity end to end. Broadband networks, on the other hand, have a variable delivered capacity, and typical router-based prioritization methods are fundamentally ineffective in delivering end-to-end QoS control.

Hughes ActiveQoS employs a number of innovations to deliver true QoS control over broadband. ActiveQoS analyzes the end-to-end broadband connection at each branch location in real time to determine the current capacity, then rate-limits traffic based on the observed capacity. By dynamically balancing ingress traffic to available network capacity on a site-by-site basis, true end-to-end prioritization is enabled, as is the delivery of real-time traffic with controlled jitter and packet loss. For example, ActiveQoS-optimized VoIP can deliver toll-quality audio (MOS 4.0) even on a congested link, and Web performance improves by a factor of four.

Enterprise broadband: Enabling the next-generation WAN

Recent major shifts in business and IT strategies have put a significant strain on data networks, especially for highly distributed organizations, such as those in retail and related industries and government public services departments. The rapid growth in the use of cloud services and mobile applications, including guest access, Wi-Fi, and video usage, are demanding high levels of service and performance availability across the WAN. Next-generation networking solutions for distributed organizations need to deliver on five major requirements:

• High capacity: Up to 3Mbps to 6Mbps for small branches, and up to 10Mbps for others
• Robust security: Premises and network-based security
• Flexible connectivity: Secure connectivity to the Internet directly from branch sites
• Application performance: Consistent, predictable performance for real-time and data applications, including emerging cloud-based services
• Reach: Delivery of next-generation capabilities to all sites, not just selected large ones

For MPLS, last-mile access has remained centered around the familiar T1 line. While T1/MPLS delivers a sound convergence platform, the rapidly evolving requirements of distributed enterprises have significantly outstripped its capabilities in terms of capacity. Ethernet access is touted as the next-access technology delivering high capacity, but its limited geographic availability and high costs make it of little real value to the IT executive for the distributed network.

Hughes enterprise broadband VPNs employ integrated performance optimization technologies to address these drawbacks, delivering fully managed and cost-effective network solutions to IT organizations in two ways:

1. By supplementing current MPLS: Deployed as a co-primary WAN and providing cost-effective, high capacity to the site. The network can also deliver backup functionality to MPLS, backing up all applications while maintaining QoS.
2. As the primary WAN: Deployed as the primary WAN, Hughes managed VPN can deliver higher capacity at savings of 30-plus percent compared to a typical MPLS network.

ActiveQoS is a key enabling technology that underpins Hughes next-generation WAN solution by delivering high application performance over broadband.

Under the hood of Hughes ActiveQoS

ActiveQoS provides QoS over broadband via three core elements:

• Network capacity determination: The ActiveQoS algorithm actively tracks available network capacity and latency on an end-to-end basis for each site in real time. ActiveQoS measures upstream and downstream capacity separately, and it can track capacity changes in fractions of a second. The algorithm is tuned to react rapidly when capacity reduction is detected, as well as when real-time traffic is present in the data stream — with two efficiency dimensions: (1) operating at a very high fraction of network capacity, and (2) introducing minimal overhead for capacity measurement.
• Traffic governance: In a typical branch, LAN bandwidth overwhelms WAN bandwidth. And most applications are bandwidth aggressive, resulting in both poor performance of other applications and diminished overall performance due to TCP’s inefficient congestion avoidance methods. By tracking actual available capacity in real time, ActiveQoS ensures that a site’s aggregate traffic offered to the broadband network — both upstream and downstream — is rate-limited to the measured available capacity; allowing real-time traffic to flow through with minimal delays, while shaping other traffic appropriately based on their prioritization. ActiveQoS also incorporates advanced TCP optimization to maximize the usable throughput of the link.
• Effective prioritization: With network capacity determination and traffic governance in place, traffic is prioritized at the edge into one of six levels. ActiveQoS includes a heuristic-based Smart Classifier, which automatically classifies traffic flows according to their QoS needs, with one real-time level and multiple data levels. The available capacity is allocated across these traffic flows, with capability to ensure minimum reservation for low priority traffic that ensures no starvation.

Multibranch, distributed organizations are increasingly adopting a host of sophisticated, bandwidth-hungry applications to increase competitiveness and improve efficiencies. Advanced performance optimization technologies like ActiveQoS are now enabling them to meet their growing voice, video, and data convergence needs using cost-effective broadband networks instead of more expensive MPLS alternatives. By transparently adding strong end-to-end QoS to commonly available, economical broadband networks, they deliver better application performance and reduced transaction response times, cost-effectively meeting the needs of these organizations.

New Tech Forum provides a means to explore and discuss emerging enterprise technology in unprecedented depth and breadth. The selection is subjective, based on our pick of the technologies we believe to be important and of greatest interest to InfoWorld readers. InfoWorld does not accept marketing collateral for publication and reserves the right to edit all contributed content. Send all enquiries to newtechforum@infoworld.com.

This article, “Making broadband business-ready,” was originally published at InfoWorld.com. For the latest business technology news, follow InfoWorld.com on Twitter.