User expectations for the application experience have never been higher. And on the backend, as IT professionals work harder and smarter to meet these expectations, application and infrastructure environments grow more complex.
Welcome to the average enterprise today.
This enterprise runs multiple distributed services, programs, and compute languages in the background, where many issues may arise. Customer experience is constantly impacted by workload placement and capacity. But with mounting application complexity, it’s becoming more and more difficult to correlate these application outcomes with your infrastructure resources.
As a result, the question arises: how do you identify which transactions are being slowed down before they become actual errors in your backend environment and harm the end user experience?
The answer lies in the infrastructure resourcing arm of our vision for AIOps with Cisco: an all-new integration with Cisco Workload Optimization Manager (CWOM). Working in tandem with AppDynamics, CWOM becomes the industry’s first and only effort to correlate the application with the infrastructure it’s running on for application-aware infrastructure optimization.
Let’s take a look.
Where Does CWOM Fit Into Our Vision for AIOps?
Our vision, the Central Nervous System for IT, comprises three distinct pillars: Visibility, Insight, and Action.
Visibility
For application-aware infrastructure optimization to work, the first thing we need is visibility into all four layers of the stack: the application and business layer, the infrastructure, the network, and security. AppDynamics provides the single pane of glass needed for that holistic visibility across public and private cloud environments.
Insight
The circle you see in the middle of the above diagram — that’s Cognition Engine, a machine learning and AI algorithm that makes correlations between the time series and events data from each layer of the technology stack, plus any alerts and APIs you’re using, to create a unified dashboard.
CWOM sits on the orchestration side, where it can take into account both infrastructure metrics (CPU, memory, data storage, etc.) and application performance (more specifically, the end-user experience) to make decisions on how to optimize your environment.
Action
Here’s where CWOM really shines — through, for example:
- Infrastructure provisioning: If AppDynamics detects business and application performance issues and the correlations between them, and CWOM correlates these issues to underlying constraints in your infrastructure, it’ll take action to rightsize your infrastructure environment for your particular needs at that time.
- Cost optimization: If AppDynamics identifies app code hotspots driving up infrastructure costs and, added to CWOM, provides insight into underutilized infrastructure that could be optimized for cost efficiency, CWOM will find the optimal intersection of app performance and infrastructure allocation so you know where IT spend can be more efficiently allocated.
The process is continuous. CWOM starts a feedback loop that confirms improvements in business and application performance, so that when the same issue is detected in the future, you already have the insight to take immediate action.
CWOM in Action
The screenshot below shows a traditional CWOM deployment without AppDynamics. It starts at the virtual machine layer and works its way down through the data center, where you can move things around to feed the resourcing ask above. For example, is the virtual machine asking for more resources from the host? If so, can we spin up more hosts, or should we spin down the virtual machine?
What AppDynamics adds is several application-centric layers up top, including the business application (in our example, AD-Financial Lite):
This is where the AppDynamics-CWOM integration strengthens what CWOM can do. When we send application data from AppDynamics into CWOM, we bring the application and end-user view into the infrastructure, making the infrastructure application-aware.
Let’s click on the business application. Here we see a different flow: the infrastructure specific to that application.
So, we can narrow down the infrastructure to the issues related specifically to AD-Financial Lite, all the way down to the data centers and disk arrays.
Next to that, we can see different policies and details of what’s happening in the application.
And if we hover over Application Server, for example, we can see critical actions to take with the application servers.
Clicking here reveals additional information: a list of application servers and actions. If we click Actions, we can see there are some issues with heap that need to be scaled up.
We can see in the last JVM that we should scale down heap because we’ve overallocated resources there.
As we go through each of these layers — Application Server, Virtual Machine, etc — we can see not only how to optimize that layer and estimated cost savings, but also the relationships between those layers and the application.
That means, with those additional application-centric layers, we can move things around without actually compromising the performance of the application running on this infrastructure.
CWOM in AppDynamics
The integration is bidirectional: You can take these actions within the AppDynamics interface, too.
Take a look below at the AppDynamics dashboard, where we’re collecting all of our business, transaction, and application/infrastructure KPIs:
Right away, in business KPIs, we can see issues with policies processed: about $300k of revenue is at risk.
On the right, we can see the data we’re pulling out of CWOM. Click Application Server or Virtual Machine, and we’ll see CWOM’s recommendations for scaling up the heap, just as we saw in the CWOM interface:
From there, we can expand each action to see more information, drill down, or execute the action, all without leaving the AppDynamics screen.
Or you can set these to automate actions and simply review the actions taken.
Either way, by making the infrastructure application-aware, we’re getting recommendations with the app and end user in mind. This ensures the infrastructure is optimized not only for efficiency and performance but for end-user performance as well.
To Sum Up
AppDynamics correlates the application to business performance.
CWOM correlates the application to infrastructure.
The result is a closed-loop model, bridging applications and infrastructure operations together like never before. From visibility to insight to action, we can see all interconnections between your application metrics and infrastructure operations. We can correlate data for a common view across your infrastructure and application teams. And by tying infrastructure resource decisions to the performance of business-critical applications, we can show IT’s value to the business.
So, what does this look like in practice?
In a recent webinar, Performance at the Speed of DevOps: I talked about how DevOps teams use CWOM to create a feedback loop that ensures continuous improvement. Click below to watch on-demand: