July 26, 2012

Make sure Hybrid Networks deliver their promise: cost effectiveness and performance increase at the same time!

(Sponsor-Contributed Paper)

Business-driven IT transformations (i.e. Unified Communications, adoption of cloud applications, the growth of Internet traffic for business and private usage at the office) imply a massive evolution for traffic delivery across the enterprise. Yearly traffic doubling is no longer an exception. Backhauling Internet traffic to the data center is too costly and adds unnecessary delays that degrade application performance.

To address these challenges, enterprises have a new option that makes perfect economic sense in today's IT environment: roll-out hybrid networks that combine high quality MPLS and low cost Internet VPN with an optional local breakout to the public Internet.

Hybrid networks mean a breakthrough in both performance increase and cost effectiveness, but they need to be managed properly for the project to yield anticipated benefits:

  • Hybrid networks' additional complexity needs to be properly managed. Traditional path selection mechanisms, such as policy-based routing, are inefficient, cumbersome and error prone;
  • Hybrid networks must be aligned to business requirements. The selection of the right network access (MPLS, Internet VPN, local Internet breakout) and the way application traffic is managed must guarantee the performance delivered to each user at all times;
  • Hybrid networks must increase business continuity - not put it at risk. This is particularly true when one of the accesses fails and the other one must handle the entire application traffic.

How Advanced Application Performance Management Meets these Challenges?

Advanced application performance management systems are able to overcome these challenges. They constantly monitor and dynamically select the best path to use among Internet and MPLS so that all available bandwidth is put to use without any risk to business critical application performance. They dynamically adjust network behaviour and resources to the exact application traffic demand. This allows business applications to run with the right performance without costly bandwidth upgrades.

To really deliver their promise, they should have the following characteristics:

  • Be unified to provide enterprises with the perfect combination: performance optimization, business continuity and IT savings;
  • Be intelligently load balanced to make sure application flow will use the best-suited network for an optimal application performance (business critical ERP applications that usually take the high-quality MPLS network can start to use the Internet if it becomes faster for any reason);
  • Be automated to not add complexity and heavy management overheads.

What is the Business Value to Control Cloud-based Applications Performance?

Implementations of Ipanema's hybrid network unification over international network situations demonstrate that 80% of the traffic chose the Internet while 20% still prefer the MPLS access. This leads to a net cost decrease of up 70% per transferred Gbyte for a hybrid [MPLS + Internet] unified network compared to full MPLS.

Let's consider a branch office with 20 employees within a larger 5,000-employee company, which operates in the professional services industry sector and has decided to deploy a hybrid network. The company has an IT budget aligned to its industry standard, i.e. $975/user/month (within which $320 is allocated to applications).

This investment must be optimized and its benefits have to be guaranteed. Hybrid network managed through an advanced application performance management will:

  1. Protect the performance of business critical applications (ERP, CRM, Unified Communications, file sharing, cloud apps, Internet, etc.) amounting to an average investment of $320/user/month;
  2. Save $15/user/month thanks to automatic WAN selection, incident prevention, better helpdesk productivity and IT governance in the complex hybrid environment;
  3. Delay MPLS bandwidth upgrades by a few years and maintain Internet access at a reasonable size for a saving of $25 to $50/user/month, typically reducing by 3 the cost per GByte.

To achieve these savings, the typical TCO for an advanced application performance management solution from Ipanema, called the nano|5 with the Dynamic WAN Selection (DWS) option, is only of $4 per user per month. That cost includes the nano, central devices, management operations as well as project overheads and aleas.

Investing $4.00/user/month in the nano5 with Dynamic WAN Selection (DWS) saves between $40 to $65/user/month while protecting applications that represent an investment of more than $300/user/month.

The Bottom Line

Not deploying the right advanced application performance management solutions in the branch jeopardizes not only the success of the hybrid network deployment efforts but also the performance of the business apps. Unsatisfied users and business managers will complain, forcing IT to upgrade the MPLS network at a high cost and to postpone the rollout of new applications (like cloud based email and collaboration or Unified Communication projects).

Because it dynamically selects the best access for each application flow, it maximizes the delivered performance and continuity, while optimizing the usage of multiple network combinations (like dual MPLS access, dual service providers or MPLS + Internet).

By making all network accesses available for use (no more passive back-up lines), Dynamic Wan Selection (DWS) enables enterprises to use all the capacity they pay for. It maximizes the performance of all applications over the unified network infrastructure (data, voice, video, interactive, etc.) and contributes to the company strategy to improve its efficiency and productivity through better collaboration and a more flexible IT management and licensing model.


1)   Can you explain more about how a hybrid topology looks at the branch office? Is there typically an intranet connection and an Internet connection at each site, or does each branch have one of the two?

The cloud is about delivering applications anytime, anywhere and on any device. This shifts the complexity of application delivery from central data centers to the enterprise network and to branch offices where users decide about traffic composition and loads.

Application delivery in branch offices becomes increasingly diverse as traffic flows from traditional enterprise private data centers mix with flows coming from the Internet (SaaS, IaaS and data centers hosting applications). Combined with the increasing use of voice, video and telepresence for branch-to-branch communications, the cloud turns branch offices from traffic spokes to focal points of enterprise communications. The days when all users’ traffic comes from a central hub are over – the branches are now the hubs.

Moreover, the overall traffic mix is increasingly dynamic. In the past, it was easy to find patterns in user behavior. A network manager could see a peak in email traffic in the morning as users arrived in the office and a peak in Internet traffic at lunchtime. Repetitive business tasks such as financial consolidation appeared at specific times during the month or the quarter.

Why use stable but relatively expensive MPLS bandwidth to carry all WAN traffic when less expensive Internet bandwidth can easily absorb growing capacity requirements? Instead, why not combine MPLS and Internet more effectively? A hybrid MPLS + Internet network makes a lot of sense for most companies and many are taking steps to a hybrid network infrastructure. Internet bandwidth was once considered only for backups of MPLS lines. It’s now time to use both links at the same time and get the best of both worlds.

Just making sure, we are assuming that in an ideal situation, the architecture would look something like this diagram - except, of course, there may be multiple datacenters and and many branch offices?

This is correct: in an ideal situation, each site will be “hybrid” with one of each network (MPLS and Internet).

2)   How does the nano go about determining which link(s) to use? What are some of the parameters that are measured in making this determination?

The nano, using the Dynamic WAN Selection option automatically chooses the best access for each application flow to maximize the delivered performance and continuity, while optimizing the usage of each available network at the same time. WAN selection criteria include quality parameters like network delay, jitter and loss, as well as the available bandwidth. Evaluated in real-time, these criteria are weighted according to the corresponding Application Performance Objectives (APOs) set in SALSA: voice will usually look for the fastest path while email might prefer the largest available one.

3)   I note that in two of your prior papers (Unified Communications and WAN Performance - A Delicate Balance and Cloud Applications and WAN Performance - Another Delicate Balance), the estimated price per user per month is about $3.00. But in this paper, with the Dynamic WAN Selection option, the price is $4,00 per month. If a company starts with a more basic model, can this option be added, or is new hardware needed?

No need for new hardware, the system is scalable and options can be added when the enterprise need it.

Search Webtorials

Get E-News and Notices via Email




I accept Webtorials' Terms and Conditions.

Trending Discussions

See more discussions...

Featured Sponsor Microsites



Please note: By downloading this information, you acknowledge that the sponsor(s) of this information may contact you, providing that they give you the option of opting out of further communications from them concerning this information.  Also, by your downloading this information, you agree that the information is for your personal use only and that this information may not be retransmitted to others or reposted on another web site.  Continuing past this point indicates your acceptance of our terms of use as specified at Terms of Use.

Webtorial® is a registered servicemark of Distributed Networking Associates. The Webtorial logo is a servicemark of Distributed Networking Associates. Copyright 1999-2018, Distributed Networking Associates, Inc.