Optimization
This section on Optimization describes the features that yield the Ipanema System's unique ability to guarantee the performance of critical applications under all circumstances.
Optimization features allow to:
Define Application Performance Objectives per user and enforce them globally over the WAN
The System's unique ability to guarantee application performance results from its objective-based approach to traffic management. The basic principle of objective-based traffic management is to define what resources the network should deliver to each end-user application flow. This is in contrast to the traditional approaches which try to manually configure how traffic should be managed and hope good application performance will be the end result. In our objective-based approach, the means to achieve the application performance objectives are computed automatically in a dynamic and global fashion to provide a second-by-second updated, optimal set of traffic management parameters.
With Ipanema, all enterprises need to define are which applications matter the most for them and what are the criticalities for their business. A higher criticality is set for applications for which the business needs more availability. A low criticality does not mean that the application does not matter for the business. It simply means that in a time of high demand, the business can tolerate the application suffering from lower quality for some period of time.
Applications are defined in the System using a combination of layer 3 to 7 criteria. For example Citrix and HTTP applications can be defined based on published application names and URLs.
For each application, a per user service level is set. A per user service level defines what the network should deliver in terms of network resource for each user of a given application. Resources are defined using bandwidth, delay, jitter and loss thresholds. This service level information is contained in a dictionary in the central software and can be customized to match each enterprise's unique application requirements.
Once the objectives have been defined, the Ipanema System is able to first assess whether or not the objectives are currently being delivered over the network. In this example, site 4 has serious application performance problems, including high-criticality applications such as SAP.
When Ipanema Optimization is activated, the System automatically reorganizes available resources to optimize end-user quality of experience and global productivity over the WAN. In this example, at site 4 the end-users of critical applications are now served with the required network resources, resulting in the users of critical applications being 100% satisfied across the entire network.
The vastly improved allocation of the exact same network resources enabled by Application Performance Objectives results in overall user satisfaction jumping from 57% to 78%. Only the users that are still lacking network resources even post Optimization will require additional efforts to improve their application performance typically a bandwidth upgrade.
Guarantee the performance of critical applications under the toughest conditions
Thanks to its dynamic and global nature, the Ipanema System can guarantee application performance better than any other solution on the market.
The two charts below are from tests that were conducted by the Tolly Group to measure the ability of the System to deliver good end-user quality of experience in real-world conditions. The test environment was based on a network with three sites and a mix of multiple business applications. The test scenarios simulated the continual changes in the number of end-users of different applications, which is found in any enterprise network.
The tests results demonstrated our unique ability to guarantee the performance of critical applications under the toughest conditions.
Globally manage meshed flows with Cooperative Tele-Optimization
In modern networks, competition for network resources by different applications occurs not only among applications but also among sites. In this example, branch office C accesses a critical SAP application over the WAN from the main data center B. At the same time, the same user’s email application synchronizes data over the WAN from regional data center A. Then a second user in the same branch C also generates email traffic from B. The resulting competition between application flows from site A and B creates congestion at the PE router at site C and impairs the performance of the critical application. Although such competition can be handled with a traffic management device located in site C (as long as the traffic from site A and B does not contain non-TCP traffic), controlling it on the destination side is not optimal.
Through global management of the network traffic made possible by Cooperative Tele-Optimization among the devices located at the data centers, the congestion in the PE router at C can be completely avoided, even without an ip|engine in C. The cooperating ip|engines in A and B exchange information in real time about the flows they are controlling, and from that they detect that they are both sending traffic toward C. They dynamically compute the bandwidth that should be given to each user session going to C based on their shared knowledge of the traffic mix, its business criticality and the resources available at C. They thus effectively prevent any congestion from happening in the destination PE router by controlling the traffic at the source before it ever enters the cloud in the first place.
Cooperative Tele-Optimization dramatically reduces the costs and the complexity of application traffic management over large distributed networks by removing the need for appliances in branches.
Dynamically protect interactive applications and enable voice/video/data convergence with Smart Packet Forwarding
The Ipanema System optimizes more than just bandwidth – it also handles the other characteristics of network resources: delay, loss and jitter. Application flows are forwarded differently depending on their characteristics and on individual user behavior through the Smart Packet Forwarding technology.
Packets belonging to real-time flows, such as with VoIP or video, are forwarded so that no unwanted jitter, delay or losses are introduced.
Packets belonging to data-transfer flows, such as with FTP or email, are forwarded so that they receive the appropriate bandwidth resources, without degrading other delay sensitive flows.
Packets belonging to interactive flows, such as with Citrix or Windows RDP, typically require the lowest transit delay. They are analyzed by Ipanema to detect what the user behavior is at each moment of a session. This unique technology is required to cope properly with “hybrid” flows such as when users load or save a local file inside a Citrix session aimed at document editing. Ipanema Smart Packet Forwarding prevents the data-transfer phases of the interactive flow from freezing up other user’s cursor in a parallel Citrix session.
Automatically select the best access link with Smart Path
Smart Path is an additional feature to Ipanema devices deployed in the branches. Its aim is to combine the capacities of multiple links, typically a primary business grade link and a cheaper Internet DSL line to increase both quality of experience and business continuity. Smart Path leverages the objective based nature of Ipanema System to choose which path each application flow should take. Smart Path relies on the Visibility and Optimization features of the Ipanema System to take dynamic decision based on the real-time status of the links (available bandwidth, quality …). Each path selection is performed according to a comparison of this status versus global Application Performance Objective and the global analysis of the traffic mix. On top of per flow network resource requirements (bandwidth, maximum loss rate, jitter, delay …) and Criticality, Application Performance Objectives include an additional parameter: the Sensitivity of the applications that is used in conjunction with a security parameter for each link.
Smart Path intelligently balances each business critical flow of the branch on the most appropriate link. Less critical applications typically run on the cheaper lines. When the traffic mix makes it possible, the less critical traffic is able to use the available bandwidth of the primary link. When the primary bandwidth is exhausted and if the applications sensitivities allow it, the bandwidth of the secondary link can be used to absorb a peak of demand from users of the critical applications.
