01/03/2017
Here’s What No One Tells You About Application Performance
An increasing number of businesses are relying on mission-critical applications for better efficiency, collaboration, cost savings, and revenue generation. The ugly side of dependence on such applications is the negative impact they can have on the business when there are outages or severe performance degradations.
With mission-critical time-sensitive applications rapidly migrating to the cloud and becoming SaaS-based, it’s becoming nearly impossible for IT to control and deliver consistently high performance for such application for offices and users worldwide.
Let’s examine what causes performance degradations for mission-critical applications and what are the best practices of addressing them.
It’s easy to blame the network when your apps get slow, but there’s more to the story than simply high latency.
Is MPLS an Issue for Applications?
Recently, a Network World Contributor reported on a test regarding the difference in latency between MPLS connections and an SD-WAN network, to see which performed better.
The experiment involved a manufacturing company whose data center was hosted in Los Angeles and had 17 global offices around the Pacific Rim. They were using several mission-critical cloud- and SaaS-based applications, such as SAP and Common Internet File System (CIFS), as well as hosted voice and video.
The initial assessment discovered there was a latency fluctuation as high as 100ms with their MPLS connections. From the test, they concluded that, in certain locations, SD-WAN combined with broadband links performed better than MPLS for application performance based on latency alone.
Reducing Latency Alone Cannot Solve Application Performance Issues
What the study failed to cover is that you must factor in more than latency when it comes to application performance. Just because you have low latency between two points using an edge-based SD-WAN device, doesn’t mean the application performance will be any better.
Why?
It is possible for a network to have low-measured latency, but still experience a high amount of packet loss and jitter. This will result in a choppy application experience for the end-user, and could even result in a greater lack of productivity. If you want to achieve optimal application performance for end-users around the world, your enterprise network needs to ensure that the quality of data delivery is seamless and consistent.
Packet Loss: The True Villain
It’s true that the amount of time it takes for data to traverse the full distance of the network depends on latency, but how quickly your applications respond depends on both latency and packet loss.
Perhaps you solve your latency problem, but still have a highly congested network and a long distance over which you’re sending data, such as across continents. You’re backhauling your MPLS through the data center to connect to the public Internet so your network can connect with cloud- and SaaS-based applications, and users depend on VoIP and web conferencing to stay connecting between global offices.
As routers become overwhelmed with data, they drop packets or send them back. So even though the actual speed of the network hasn’t changed, the packet loss has – which means that application performance will still suffer, as they wait for inconsistent packets to arrive, especially across large distances.
Solving the Packet Loss Problem
Most people first look to reduce packet loss by increasing bandwidth. However, this does little to improve application performance, as bandwidth is still shared on the middle mile of the public Internet. Therefore, heavy packet loss is still a reality when sending data over large distances.
The best way to improve bandwidth usage in this situation is with WAN Optimization. When traffic is optimized for improved throughput, the result is a more consistent and reliable application performance.
However, if you’re using MPLS, that means an additional investment in WAN Optimization technology – which means more hardware and services that you must manage. Some organizations solve for this with additional investments on top of their MPLS and WAN Optimization, like SD-WAN. However, this also means more time, money, hardware, and service.
The best way to ensure good application performance without the additional investments is by investing in a network that provides stable connectivity at the outset, i.e. a combination of consistent latency and negligible packet loss.
Infoquick does just that with our global private network that provides seamless connectivity to branch offices as well as cloud and SaaS applications. In addition, it comes with layered with WAN Optimization technology to improve throughput and reduce investments in bandwidth.
Infoquick customers have recognized reducing packet loss as a key determinant of faster application performance.
In fact, this increased optimization of bandwidth and throughput leads to a greater costs savings in connectivity not found with legacy networks or Hybrid WANs.
It’s important for all IT buyers to keep this packet loss question in mind when evaluating various solutions based on performance measures. This way, you have a more complete picture of the kind of network performance you can expect from your investments – and just how much you’ll truly have to invest to get the performance you need.
Is your enterprise looking to improve application performance worldwide?
