So you’ve decided to take the infrastructure as a service (IaaS) approach to cloud migration. That’s great—you’re on your way to realizing the cost savings and flexibility of cloud computing.
But the decisions don’t stop when you choose between SaaS, PaaS, and IaaS. Choosing your public cloud provider—Amazon Web Services (AWS), Microsoft Azure, or Google Cloud Platform (GCP)—is the next step.
Each public cloud provider requires unique considerations, but you should have one goal in mind, regardless of the provider: continuous public cloud monitoring.
Monitoring Considerations for Amazon Web Services
AWS is loaded with services to give you flexibility in application migration. However, this leaves you with choices as to which ones will work best for your specific business needs.
The following two choices could negatively impact end-user experience if not handled properly:
- AWS Lambda vs. EC2 for Computing: This decision is similar to the one you make regarding SaaS, PaaS and IaaS. While Lambda can take administrative tasks like provisioning capacity, monitoring fleet health and security patching out of your hands, EC2 offers the freedom and flexibility to support customized applications. Code deployment and monitoring are critical to maintain the user experience. But even if you choose Lambda to offload admin tasks, you have to remain proactive with continuous monitoring.
- Amazon S3 vs. Glacier for Storage: Storage is a critical component of application design if you want to maintain good end-user experience in the cloud. S3 and Glacier are often discussed in the context of backup, but the use cases for these services are nearly endless. The goal of each is to minimize latency and make application availability seamless for users. Amazon S3 is a robust storage system for mission-critical applications whereas Glacier is meant as a “write once, retrieve rarely/never” service that minimizes costs.
There are monitoring capabilities that come built into many AWS services, but modern end-user experience demands require more than those traditional BGP approaches.
NoSQL vs. SQL Databases in Microsoft Azure
The type of database you choose for your cloud application is critical for performance and end-user experience. SQL has been the traditional go-to database for decades, but modern demands are causing developers to look elsewhere. SQL databases just can’t keep up with high volume, velocities and varieties of data in some use cases.
NoSQL databases have emerged to enable storage of unstructured data at scale. For use cases like social and Internet of Things applications where document databases and key value stores are common, NoSQL will facilitate better performance.
Microsoft Azure offers multiple services for both NoSQL and SQL needs, including Azure Redis Cache for NoSQL and Azure SQL server. When you’ve decided which type of database is right for your application, you have to set yourself up for success by implementing continuous monitoring.
Maintaining End-User Experience in Google Cloud Platform
Google Compute and BigQuery are major components of the Google Cloud Platform. Google Compute promises unmatched virtual machine performance to maximize your workloads. However, GCP doesn’t highlight its inherent monitoring capabilities. If you want to ensure the end-user experience consistently matches GCP workload performance, you’ll need continuous public cloud monitoring.
In addition to Google Compute, BigQuery works on the back end to help power your data analytics processes—but most importantly, comes with cloud monitoring functions to help track workloads. However, workload performance is just one piece of the puzzle. You need continuous public cloud monitoring to gain insights from the end user’s perspective to maximize the ROI of your GCP investment.
Continuous Monitoring: The Link Between All Public Cloud Providers
There are many factors that go into a decision between public cloud providers. But no matter what, the end user has to be your top priority. No matter how much money you save operating in the cloud, you’ll never see ROI if end users can’t actually access and make the most of cloud-hosted applications.
With continuous monitoring, you can gain visibility into every aspect of application performance and the end-user experience. Your infrastructure may be hosted in the cloud, but that doesn’t mean you offload all responsibility for maintenance.
If you want to learn more about what exactly you should be measuring in the cloud, download our free guide, The 5 Network Metrics You Should Keep to See into the Cloud.