Scalability – Latency and performance

The scalability of the SaaS platform is a key consideration for business applications. Being able to scale out and scale up to support seasonal workloads or spikes in user activity will impact the overall user experience (both staff and customers) and effectiveness of business processes.

The cloud scalability parameters available in SaaS differ from a traditional on-premises system. Instead of adding more servers or increasing the power of machines, these parameters could be translated as available API capacity. You also shouldn’t assume that cloud means infinite scale and computing power that can process everything thrown at it. The good old coding and design best practices are still relevant, but might need to be adapted. Although managing and scaling cloud services is complex, increasing or decreasing your data storage, processing power, and networking requirements can be done seamlessly. In many cases, you can do so automatically or with simple configuration changes. This microservices architecture, including capacity-based routing to resources and storage management, is transparent to Microsoft customers.

To summarize, the SaaS cloud offers the flexibility to scale horizontally and vertically to support thousands of concurrent users. Embracing the capacity-based model for resource consumption in the cloud helps not only build optimized applications but also a better plan for operating cost post deployment.

Performance is a key consideration for business applications—it not only impacts the end user experience and adoption, but can directly impact
business goals and key performance indicators (KPIs) of success.

In on-premises deployments, enterprises had complete ownership and control of the infrastructure, and could guarantee and monitor applications’ latency and performance. In the cloud world, it’s not as straightforward.

Research shows that a few milliseconds o f latency lead to a big percentage drop in page load times. For e-commer ce companies, this could mean a sizable drop in user attention and therefore sales revenue. Low latency is not a good to have, but a critical deciding factor in an enterprise’s brand and r eputation. The same applies to back office business applications operating in the cloud—user experience and productivity can be significantly impacted in a high-latency environment.

Several contributing factors can impact application performance, including network latency, user device and browser, application design, and customizations. SaaS services are inherently scalable, have vast compute power, and are available from multiple locations around the world, but that doesn’t necessarily guarantee performance if the solution is poorly designed, or if users are accessing the service from environments with high network latency.

Your cloud applications are powered by hundreds of pieces of virtualized hardware running on Azure datacenters sprinkled around the world. A crucial decision you must make, then, is choosing the datacenter to deploy your application so that all users of the application get an acceptable level of latency.

Performance also can suffer because of a poorly designed application. If you’re changing standard code in SaaS or deploying custom applications in platform as a service (PaaS) or infrastructure as a service (IaaS), your team must design a high-performing application and carry out proper performance tests, and only then do you deploy to production. Now, most cloud providers work on a shared network infrastructure in which your requests go through the internet and are used and shared by everyone.

Network latency is a crucial factor to consider alongside other architectural and design decisions that impact performance.

One option for businesses looking for more reliability and security is to use a private connection. Cloud providers offer dedicated channels, for example Azure ExpressRoute. These connections can offer more reliability, consistent latencies, and higher security than typical connections over the internet.