Cloud computing offers new concepts, functionality and ways of working; BUT many of the terms are confusingly similar to those you may have used in traditional IT environments.
As you explore cloud computing you will likely encounter these terms.
- Containers are basically virtual machines, just one step higher — they’re all running the same operating system. They’re faster to spin up and down (and faster to destroy after they’re done) than virtual machines. This makes them much more efficient while requiring fewer resources at the same time.
- Clusters are, of course, virtual clusters they follow the same rules as physical clusters, and they’re housed within a portion of a data center that’s been assigned to you. In a physical cluster, you’re limited to the machines you have on-premises, and if one machine fails, there can be ramifications on the rest of the system. Virtual clusters can be made of physical or virtual machines situated anywhere in the world, they can be spun up and down on demand, and aren’t likely to be brought down by the failure of any one individual component.
- Load balancing provides automatic oversight over every instance of your production environment and makes sure no servers are receiving an undue amount of strain or reaching their capacity. This isn’t really possible on-premises; the increased latency and lack of scale wouldn’t be worth the investment.
- Servers in a cloud context typically mean “space rented on a server”. Renting space on a server saves that resource for you until you say otherwise; you can scale up and down on demand, but there’s always capacity ready for you on a moment’s notice.
- Serverless is a cloud computing service model where functions of code are the unit of deployment. There are no machines, VMs or containers to manage. Computing power only exists for you when you need it; there are no “reserved” resources for you; code is dynamically run as needed and then destroyed when done. Since there are no VMs or even containers to manage, serverless computing has minimal operational overhead, making it the ideal platform for developers dev and test environments. Serverless tends to be more affordable for infrequently or sporadically used applications as you only pay for code as it runs, but it is slower than traditional cloud services since you don’t have reserved resources dedicated to your applications. For these reasons applications in continuous use or that are time sensitive are better served by more traditional cloud service models.
- Storage in a traditional data center, whether databases or files, means having a physical computer with an operating system and a configuration to run your files. Storage on the cloud is fundamentally the same, but the resources aren’t usually located in the same place physically. The code might be somewhere in a place optimized to run code, the hardware might be somewhere else optimized to run hardware, and the operating system might be in a third place optimized to run operating systems.
- Virtualization is effectively the same thing as having virtual machines (i.e., multiple instances of the same operating system on a physical computer). Instead of running these virtual machines yourself, however, your CSP will typically provide the virtualization for you. (Note: Software licensing can become sticky in virtualization, because providers may charge you based on how many virtual instances you have instead of how many physical computers you have — since you don’t own those computers yourself anymore.)
