In This Article

copy link

It is becoming more and more critical for enterprises to formulate a strategy around visibility and actionable cloud analytics.

Today, organizations can store data in a way that best meets their business needs, scaling up or down as required by demand. With a shift to cloud, there are no longer limitations to how much data can be stored and who can access this data. This is why it is becoming more and more critical for enterprises to formulate a strategy around visibility and analytics.

With the adoption of the public cloud and/or virtual overlay technologies, we are seeing solutions that introduce greater and greater separation of control from data. Although this does provide an opportunity to simplify execution of application workloads through automation, it can introduce a lack of visibility when attempting to root cause a problem in the expansive network underlay we call the Internet.

When enterprise IT applications are deployed in public cloud environments provided by companies such as Amazon (AWS), Google (GCP) or Microsoft (Azure), vendors do not typically provide the tools necessary to diagnose an Internet connectivity problem that may negatively impact the application. To quickly resolve for each scenario, IT needs new solutions that provide insight into the end-to-end topology that will identify the traffic patterns from access to each of the multicloud provider network integration points.

Cloud capabilities

Enterprises on a cloud analytics adoption journey are looking for frictionless ways to not only develop, deploy and manage their applications and services that support them, but add the ability to monitor application and network health to react in an automated way. There are three top business drivers that most enterprises envision around adoption of cloud analytics:

  1. Enhancing business process
  2. Improving the customer experience
  3. Better overall collaboration between teams

Adopting these technologies and shifting the mindset do not come without obstacles as it relates to analytics. Security (compliance) and integration (migration) are key considerations that drive towards implementing a robust solution. You cannot protect what you cannot see. By correlating the data necessary to make those "proactive" changes to policy, the end result could mean mitigation of a problem prior to it actually happening, which can reduce cost and boost overall performance.

This article addresses the key attributes of an actionable analytics solution that are required to be successful.

Beyond the cloud

Beyond just delivering applications and filtering content to multicloud, IT organizations must consider those solutions that provide a holistic view of the applications' health. Application-aware solutions manage traffic, which not only provide visibility into segmented network latency and throughput, but also manage the applications per user-transacted analytics and logging analysis. This can give the administrator the means to isolate a potential issue, either in the network or the application (some folks refer to this is as Mean Time to Identify, or MTTI*).

For example, actionable analytics grant the ability to proactively measure risk through an application score card that will give the administrator the probability of an issue, providing user usage profile and protocol data necessary to prioritize a fix prior to occurrence.

This results in an outcome that provides the efficiency necessary to optimize the network or application servers and could change the mindset of the operation team to think more about how the network can impact the application.

Adapting for scalability

Another differentiator is the ability to send alerts when classified application usage changes. As your company grows and application capability elasticity scales, the solution should scale with demand. Identifying those changes allows the operations team to trend over a period of time, hopefully implementing scale-out and scale-in monitors that will drive resource utilization based on application demand.

Above, we touched on the policy which is the basis for "intent-based-networking" currently being discussed as high-level in the industry. Extending the solution even further is the key to closing the loop to a true actionable analytics solution. This can be done by leveraging machine learning, cognitive computing and deep analytics capabilities to provide greater levels of programability, automation and security integration, while reducing time spent on manual network configuration and management.

Constantly adapting based on a system where end-users dictate how the application and network should perform and correlate to analytics provided by data learned from usage will result in application service adaptation to effectively provide an optimal service level.

Let me know your thoughts on implementing cloud analytics in the comment section below and take a look at the first article in this series, Building Network Automation into Your Organization.

*MTTI – Mean Time to Identify – The time it takes an organization to isolate a potential issue in the network or application that could negatively affect overall performance. For example, MTTI could be greatly reduced by isolating response times from each application servers as compared to network round trip latency.