How do IT staff manage user application experience in a complex environment?

As BYOD, wireless, unified communications (UC) and cloud-based SaaS applications gain traction, new user application analytics technologies can help IT manage the quality of the user experience.

Understanding how end users access and experience network applications requires a new approach. 10 years ago, many enterprises may have relied on Intel-based Windows devices connected to Ethernet to access local applications. This can improve the end-user experience in existing use cases. However, this scenario is no longer the norm.

Today, most applications are supported by a dynamic infrastructure. This infrastructure is distributed across on-premises environments and public clouds. Client devices are diverse in terms of hardware, operating systems, drivers, and more. These devices access the network wirelessly. The dynamic RF environment and potentially unreliable Internet connections undoubtedly increase the overall complexity of the IT environment.

What insights can traditional application performance monitoring (APM) and related technologies provide in this new landscape? APM tools answer the question: Once connected to a Web server, how long does it take for a client transaction to be satisfactorily serviced? This takes into account transaction times between Web servers, application servers, database servers, etc. APM focuses on the inner workings of multi-tier applications in a data center.

However, if IT managers think about the applications that exist in the enterprise today, they will find that this portion currently only accounts for a small portion of the overall user application experience.

First, IT staff are increasingly dealing with business-critical applications that are not under their control. When using third-party applications, IT staff cannot monitor the third-party data center, which makes APM technology meaningless. Second, using APM presupposes that the client can effectively connect to the access network and therefore the application Web server. If this connection fails, APM tools are powerless to collect or analyze data from the wireless LAN infrastructure and campus network services, nor can they analyze the health of the WAN Internet link.

Because of its limited scope, APM tools are simply not suited to handle the complexity of the problems users face before transactions even connect to the application Web server.

What's missing?

To get the most insight into the user application experience, network operators need insight into the client device's ability to: connect to the wireless access point in a timely manner, authenticate via RADIUS, obtain an IP address via DHCP, resolve domain names via DNS, and finally send and receive data from the Internet. If any step in this process fails, thinking about "application performance" is meaningless because the user cannot even access the application.

With network managers having to sift through mountains of network, application and user data from numerous silos, gaining a comprehensive view of all of these user network transactions can be challenging.

The changing dependencies between networks, devices, and applications require a more holistic approach to IT analytics. Performance metrics for applications, wireless radios, network services, and Internet links, as well as device type and OS information, must be collected simultaneously and comprehensively. Next, these metrics need to be analyzed and correlated across time and other dimensions, such as location, SSID, and VLAN. Finally, advanced machine learning algorithms need to be applied to these metrics to proactively discover insights and remediation recommendations.

Network managers are primarily responsible for operating wired and wireless infrastructure, but as cloud-based SaaS applications, smart devices, unified communications and wireless become more prevalent, a broader, client-based perspective is needed to answer fundamental questions about the end-user experience.

User application analysis technology debuts

Gaining a more complete understanding of enterprise application behavior from the user's perspective requires the simultaneous collection and correlation of customer wireless performance metrics, network service performance metrics, device OS information, and, most importantly, application performance metrics. This is most seamlessly accomplished through a combination of real-time packet analysis and data collection from other enterprise systems, including wireless LAN controllers, and APIs from unified communications and other applications. Finally, advanced analytics, including machine learning, must be performed on the data to automatically uncover insights.

New User Application Analytics (UAA) software technology is expected to provide just this broad reach. UAA software runs passively within an enterprise access or campus network and requires no client software, comprehensive testing, additional sensors or infrastructure hardware.

In addition, UAA software uses real-time application identification technology for every customer transaction that occurs on the network, performs deep packet inspection on wired packets, and analyzes wireless metrics from the wireless LAN controller. By looking at the interaction and response time of the underlying protocols (such as HTTP, TCP, and RTP, etc.), applications can be identified in real time and performance metrics can be measured in real time.

You can then combine many different dimensions to identify and compare application behavior for a user, user group, discrete application, location, or even a specific virtual wireless network (SSID) to get real answers to questions about the actual end-user application experience, such as:

• What is the baseline user experience for application X?
• When an app’s user experience is poor, is it the wireless, the device, the internet connection, or the app itself?
• How will the user experience vary due to different locations, subnets, SSIDs, and RF bands?
• Can users reliably connect to the wireless? Are link layer services being provided correctly?
• If I increase the capacity of my DNS servers, how many customer hours of SaaS user experience issues can I save?

With UAA software, IT can easily identify application adoption trends, traffic usage, individual and systemic customer events and root causes, historical trends and overall quality of experience for any user at any time.

Unified communications monitoring is one of the top use cases for user application analytics, primarily due to the massive adoption of unified communications and the digital transformation of enterprise multimedia business content.

For unified communications applications such as Skype for Business, UAA connects to the vendor's API to get a truer picture of the user application experience. Call attributes are captured, analyzed, and correlated with other network transaction metrics. One such call attribute is the Mean Opinion Score (MOS), a standard telephony metric that describes the actual user experience.

Today's enterprise applications have become highly distributed, with components often scattered across local data centers and public clouds. For network managers, a good user application experience can no longer be isolated from the underlying wireless access infrastructure, network services, and mobile devices used to access them. This now puts IT staff in a dangerous position: they have to dig deep into a vast amount of network, customer, and application data to understand the actual user experience.

New user-application analytics technologies promise to finally give IT a comprehensive view of the network from every angle, providing hard numbers to justify the huge annual IT dollars invested so that users can work efficiently on corporate networks.

Original title: Getting a grip on user application experience, author: Anand Srinivas

[Translated by 51CTO. Please indicate the original translator and source as 51CTO.com when reprinting]