Application performance monitoring, which measures the availability and performance of your applications, is pivotal in ensuring your software is delivering its business value.
One weak link in the system can massively impact the performance of your applications, resulting in frustrated users and productivity dips. Therefore spotting poor application response time and addressing the issue quickly is crucial to both, the user experience and bottom line.
An application may be running, but this doesn’t mean you are achieving peak performance. Today’s software is complex and relies on a network of servers and infrastructure to support it. For example not every application performs at its optimum in the cloud, so it is important that the enterprises check out performance speeds before migrating applications to the cloud so they can solve issues that may arise or consider alternative applications.
According to the Manage Engine Application Performance Monitoring Survey, 32 per cent of organizations said they find out about poor application performance from their end users. This indicates that application performance testing is still not a priority for IT departments, even though these applications are often critical for business. In addition, 60 per cent of respondents reported that it can take up to four hours to address application performance issues, once they have been reported.
There are many applications that are prone to performance downturns. According to Laurent Jankowski, Application Performance Consulting Services Manager for Orange Consulting, a number of companies are suffering poor application performance from Microsoft Office 365 and Skype for Business applications, for example. These include slow response times with Office 365 and poor audio/video quality with Skype for Business.
According to Jankowski, the first step to resolve these issues is to have 360-degree real-time application visibility of the network, applications and end-user experience. Orange has developed a solution based on Riverbed SteelCentral, which provides performance insights that expose and diagnose any bottlenecks in the applications, network or infrastructure. Problem solving depends very much on server location and enterprise infrastructure.
“As the architecture of these applications is very complex, we need to work out where the servers are located for each application,” explains Jankowski. “This will determine what type of solution we use, as there are many options. For example if it is an issue on the Microsoft side we will use Riverbed to capture the traffic and troubleshoot the flow directly. Depending on the location of clients we may use wide area network (WAN) optimization to optimize the traffic between the client and the server where applications are hosted.” WAN optimization allows network traffic to be prioritized to guarantee the availability of bandwidth for business-critical applications.
Optimizing the cloud.
It’s important to recognize that the big benefits of cloud in terms of flexibility, self-service and automatic provisioning are achieved at the cost of increased application complexity. If not addressed, this can have a detrimental effect on performance.
“When organizations move applications to software-as-a-service in the cloud, they need to be aware that the application performance may not be the same as the performance that they had when they hosted the application,” explains Jankowski.
In addition, cloud services are sold on usage, so it is paramount that organizations understand the cost of running applications on-premises, in the cloud and in a hybrid scenario. Organizations can then see where applications are running and if anything needs to be changed to get the most from their IT budgets.
Unfortunately, it is often impossible for organizations to design infrastructures that will optimize application performance, quite simply because they do not have an end –to-end-view. The client communicates with the web server, which can be easily identified. But after the web portal there are multiple servers located across the world, which makes it impossible to join up the dots to create a complete picture. “Organizations can’t design optimized infrastructures because they don’t have access to the entire infrastructure as the final part belongs to providers, such as Microsoft, Akamai and Amazon,” Jankowski explains.
Improving Office 356 response times.
Jankowski has worked with two large customers recently to improve the response times for Office 365. There were two approaches available. It was either redesigning the customers’ network to provide direct internet access via a hybrid network, or deploy Orange Business Services’ VPN Galerie. The latter would provide the organization with direct access to Microsoft’s data centers. “We have agreements with multiple SaaS vendors to have direct access to their data centers. This allows us to reduce latency and provide service level agreements (SLA) on the network instead of using the internet,” explains Jankowski. “Where there is no SLA there is no guarantee of performance.”
Last year Orange integrated its acceleration service Enterprise Application Management (EAM) Riverbed into Business VPN Galerie to accelerate cloud services. EAM Riverbed is a fully-managed service from Orange Business Services that delivers application acceleration and WAN optimization.
Because the internet can play a key role in application performance, Jankowski recommends that organizations re-think their internet access and usage. Companies should make sure the internet is closer to their office locations, instead of via central access. “The closer the internet access is, the better the performance for the end user. Organizations will have to decide if they want to redesign their network to a hybrid solution, upgrade their central internet access with larger bandwidth or opt for multiple internet accesses with multiple operators,” says Jankowski.
Optimizing in-house applications.
One area where organizations can address application performance is on in-house application development projects, where issues can be addressed from the design stage.
Orange, for example, works with customers before their application goes into production on analyzing bandwidth and latency issues, to ensure it is optimized for release. If the application is already in production, Orange can work with the customer’s IT team to monitor the application’s performance to detect if it has any issues on either the network or server side.
For monitoring, Orange uses multiple solutions, including Dynatrace and Riverbed SteelCentral. If latency affects performance, the application may have to be re-designed to make it less “chatty” on the network, meaning less network-round trips, to delay data transfers.
The biggest difficulty for organizations with SaaS applications today is that customers don’t have full end-to-end performance visibility. This makes it almost impossible to get a granular analysis on application performance impact. “In the future, if everyone around the table uses the same SaaS indicators and work together, there is no reason why we couldn’t solve all the performance issues that exist,” concludes Jankowski.
Faster and more productive: Enterprise Application Management mitigates the effects of latency and bandwidth constraints to improve application response times and reduce data volumes on the WAN. Find out more here.