The domino effect of performance degradations: the challenges IT teams face in getting to the core of issues

How many tools are needed for a 360-degree view of performance? The answer is, “not as many as you think”

We have heard it all before: “It’s the network”, “It’s the server”, “It’s the router”, “It’s the service provider”, and the list goes on and on. It is like an internal ping-pong game. Who, what and where to blame? Where should the finger be pointed?

Everyone has the same goal: to make the client, your end users, happy. An IT department’s clients are the users of IT managed applications and the networks. In the end, all that truly matters is a great end-user experience. If not, tickets are opened with IT, man-hours are spent on investigation, and internal finger pointing becomes a never-ending problem.

Databases, data warehouses, remote sites, servers, users, VPNs, firewalls, routers, switches and much more are handled by the enterprises’ IT team, but is IT equipped with the proper tools to troubleshoot all these complex issues in order to find the problem?

Most of today’s applications that have a client-server communication are using Transmission Control Protocol (TCP). It first starts with a handshake between the client and the server, then moves to the data being sent in sequence. For every sequence, an acknowledgement is replied to the sender. Once the data has finished transferring, the TCP connection is closed. Clients and servers can have multiple TCP sessions running in parallel and the performance every session could be different.

What is impacting your TCP performance?

TCP is designed to be loss-less, hence the handshake, the sequencing and the acknowledgement messages. However, when a loss occurs, the sender has to retransmit the lost packet. Let us picture this event but times the number of users, times the number of servers, times the number of applications. This results in an increase of bandwidth usage on the network, which also results in affecting other applications running on the same path.

TCP connections are broken and need to be re-negotiated, the TCP reconnection cycle kicks in and more packets are lost because there is no bandwidth available. Pauses, retransmissions, reconnections, packet size reductions, congestion control mechanism, etc. – these events can go on and on, but what matters the most is that the end user is being affected by a poor performance on her device.

One or many communication layers could be impacting the performance of TCP, from the network, to the transport layer, to the session layer, etc. Having a performance breakdown of all the layers requires many tools and it makes it difficult to troubleshoot. The IT department need a unified view that allows them to view, correlate, expand, drill-down and isolate events. Having a single pane of glass solution that contains all this information is what IT teams are looking for.

Accedian Skylight helps IT teams to troubleshoot, analyze and identify the issue or issues that could cause performance degradations. From the MAC address all the way to the banner loading on the webpage, Skylight offers visibility to its users in real-time for networks and applications. The ability to quickly deploy the solution, the flexibility of using software or hardware sensors depending on needs, and the local or cloud-deployed analytics platform allows IT teams to choose the solution that works best for their hybrid environments. 

Accedian uses standard and proprietary testing protocols for active monitoring. Testing the performance of communication circuits is important because the lower layers of communication are responsible for the data transfers. This can have a direct, serious impact on the applications and the users.

How do enterprises ensure that their service provider is respecting service level agreements?

The best way to find out is to run a test. Not a simple speed test but a carrier grade test that uses sequenced, precise and accurate methods of testing.

How can the IT department find out if the performance degradation is due to the DNS server, or to database queries? Working blindly is the worst way and Skylight brings visibility to IT teams, enabling them to make decisions and calculated updates and changes.

Using the key performance indicators (KPIs) that Skylight provides, a performance dashboard can be built to combine network and application KPIs to allow for expedient investigations.

Accedian offers the whole performance monitoring suite – from KPI generation and calculation, all the way to smart dashboard creation, Skylight is a perfect fit for enterprises and communications service providers that are looking to improve both quality of service and quality of experience for their end users.