This vendor-written tech primer has been edited by Network World to eliminate product promotion, but readers should note it will likely favor the submitter's approach.
With the increasing use of Web applications from mobile devices and the variety of devices accessing applications through different types of networks, a new approach to application delivery optimization is required.
Application Delivery Controllers (ADCs) at the edge of the data center can play a key role addressing this challenge, with their ability to apply contextual application delivery optimization based on a wide range of information accessible to the ADC in real time. This includes the type of device, size of its screen and the access network type.
[ HOW-TO: Shop for Application Delivery Controllers ]
We are all experiencing how mobile is changing our lives -- not only in terms of the content we consume but also when it comes to how, where and when we access the Internet. Compared to the old paradigm of Internet access through a desktop from designated locations (at work or at home), smartphones and tablets have made the Internet accessible to an ever-increasing number of users, many of whom own and use more than one device at the same time. With Internet access available practically anywhere and everywhere, it's no surprise the amount of time we spend online has grown exponentially. [Also see: "Accessing the Internet by mobile device doubled in 2011, data shows"]
Mobile trends are continuously evolving. More personal devices, like car accessories and gaming devices, now connect to the Internet, and with the strong adaptation of LTE, mobile networks are providing faster service with higher bandwidth. However, the reality of users accessing more content from multiple locations/devices at any time, has created a number of new challenges that must be addressed:
- Mobile service adaptation per device -- Gone are the days where it was enough to design your Web service optimized for just an 800x600 screen resolution. Today's Web applications and services must support different devices with different screen sizes (Smartphone/tablet/desktop), different user interfaces (touch screen/mouse/keyboard) and use different access networks (3G/LTE/Wi-Fi/LAN).
- New capacity requirements -- The increased number of users also creates large variations between average usage load and peak time usage load, which also varies across the different Web applications and services. This creates a need for contained yet flexible application delivery infrastructure capacity, which ensures each application receives the resources it requires when they are needed.
- Sustaining high Quality of Experience (QoE) -- This is now a critical online business requirement, which can only be achieved if you can actually measure and monitor user experience from end-to-end.
Contextual application delivery optimization
In order to optimize the delivery of application/content according to the device and where the request came from, it's necessary that the application infrastructure is able to recognize several characteristics of the originating client and user profile, including:
- Type of device -- This is the most basic parameter that determines which server will provide the reply. From smartphones and tablets to traditional PCs, each device may be served with a different representation of the Web application.
- Size of screen -- This can indicate the size of the objects in the page and the amount of compression that can be applied to what is being viewed.
- Access network type -- Is yet another important parameter that helps determine the best IP protocol optimization method to yield the best performance improvement. Networks with longer delays (such as mobile networks) require different optimization than wired high-speed networks.
All of these parameters can also take part in the security policy applied to the information delivered to the end user. This is especially the case in enterprise applications where an organization may have one security policy when accessing the Web application from within the organization, and a different security policy when connecting through a mobile network.
But optimized application delivery can't be achieved solely by the application itself. Smart ADCs, with access capabilities to information at the application layer, can collect all the required information in order to contextually decide which server can most optimally answer the request. In addition, ADCs can also help determine the best content optimization that can be applied (without compromising on QoE) and the best protocol optimization method for the request/device in question. Many mobile operators are already using ADCs in the core of their network to apply different traffic policies and security -- on a per-user basis.
More advanced ADCs will also be able to apply additional acceleration methods, such as smart caching, compression, transport-level optimization (e.g., Hybla algorithm) and even perform invisible content modifications, such as optimized reordering of a page's objects. This works to enable faster application response time asymmetrically -- without installing anything on the user side. That saves Web development efforts and allows unification of Web applications across all browsers, rather than pending on customer adoption of mobile apps.
Any application delivery optimization done by an ADC will not only provide better application performance but also provide valuable server offloading. ADCs with optimized hardware for Secure Socket Link (SSL) termination and content compression can effectively offload those tasks from the application server. In addition, ADCs supporting smart caching can extend the amount of cacheable content, eliminating many of the requests from even reaching the servers.
Overall, the result can be significant server offload, enabling optimization on the cost of the server infrastructure per user transaction as well.
Maintaining high quality of experience -- at all times
While mobile Web application access increases usage, it also potentially puts QoE at risk. For Web applications that drive sales (like online shops, ordering systems, etc.), any drop in performance means lower customer satisfaction, lower conversion rates and lower revenues. Oddly enough, a study from Equation Research showed that users accessing Web services from their mobile device expect faster response time. This is why it becomes even more critical that Web applications servicing mobile users will continuously maintain a consistently high QoE.
A fundamental requirement to achieve this is to have visibility on the end user's QoE with an Application Performance Monitoring (APM) tool. This tool can indicate when users are experiencing slow response times or receiving error messages in real-time. It is critical that this type of tool not only measures the performance of the Web servers but also the time it takes for the end user to receive all requested content.
A good APM tool should provide performance drop detection and the ability to perform quick root cause analysis in order to determine whether the performance drop is occurring in the application servers, the network or even the specific access network. The tool needs to be able to assess if the problem lies with a specific mobile operator's network or with a specific type of user device.
The ADC provides a bridge between client requests and server replies. As a result, it is the most natural place to collect detailed performance information, oversee server performance and network delays as well as client rendering time. Ultimately, as ADCs become smarter, the performance information collected by the ADC can also be used to further fine-tune optimization methods applied per end user device type and even per end user.
With the exponential growth of mobile Internet access together with the increasing usage of Web applications through mobile devices, application delivery optimization is becoming a more critical part of any Web application or service. The variety of devices with different characteristics equipped to access the Web from different types of networks, requires per case application delivery optimization.
ADCs, with their ability to extract all the performance information required, as well as the capacity to apply contextual application delivery optimization coupled with integrated APM capabilities, have become an essential part of any application infrastructure that services the mobile client.