Moving to Network Analysis in a Cloud Environment
Whether you are at a conference or reading the news, it seems like you can’t escape a conversation without a mention of cloud computing. No matter what the topic, from developers’ ability to deploy applications more easily on the infrastructure layer or embedding BI tools on the cloud, the cloud is seemingly affecting everyone in IT. But, what has been missing from the discussion, or so rarely discussed, is how cloud computing affects the monitoring and management of your network.
It doesn’t matter if your server is located in your data center or in the cloud, the only thing that’s changed from a network perspective is that the blame has shifted; there is still a need to monitor and analyze your network traffic. Now when you discover a problem is occurring on the application side, you have to deal with multiple external vendors, which can be a headache. Some of this can be averted though. And it first starts with the initial shift to the cloud.
A successful transition to the cloud will be one where the organization has a clear understanding of their current network. This includes establishing and understanding network performance baselines, especially for the performance of critical network traffic and applications for hourly, daily, and weekly cycles. Special attention should be given to verifying the performance of transactions that interact with multiple applications, since this is likely to be a challenging area once the applications are moved to the cloud. Other metrics to consider include packet size distribution, latency, and protocol and node usage over time. Knowing these cycles will help you determine your application and network utilization. This information is critical in determining exactly what is required for a cloud computing vendor to maintain, and hopefully improve, the performance that users currently experience.
Remember that shifting application servers to another location does not always make the most sense, and before you make the shift, keep in mind two major issues you may face: latency and security. Latency typically increases when you move to the cloud since the distance to application and data servers can increase greatly, along with the number of hops for data packets. And with the move to the cloud you lose a lot of control in improving latency.
As for security, make sure that you clearly understand your cloud provider’s security policies. The cloud vendor is storing and processing highly sensitive data and applications, so the stakes are high. It is imperative to clearly understand the cloud vendor’s security policies and procedures before contracting its services, and any appropriate verification testing should be implemented within the organization to ensure these policies are not violated.
Once your applications are in the cloud, monitoring and analyzing your network will help you compare against performance claims from cloud vendors. You can verify whether your cloud vendor is living up to their service level agreements, and act as a sort of ‘watchdog’ if they are not. You’ll be shifting from managing your own infrastructure to managing service availability and performance.
For more information on managing your applications on the cloud, check out this five-minute video:
