Over the years the public cloud has grown tremendously to a point where even late adopters are rushing to store their data and applications on the cloud, from single-run startups to a huge chunk of fortune 500 companies.
Some legacy applications cant be stored on the public cloud and need to be stored on premises because of data residency requirements or government restrictions, however, most of the data that you access on a daily basis from social media sites to streaming of online content is all stored on the public cloud.
The public cloud has far greater benefits as compared to on-premises use such as.
- High availability- all the major public clouds have divided their public cloud into regions from GCP, AWS, Azure, IBM Cloud and even Alibaba cloud. These regions are spread around the world to ensure that resources are available from any point of the globe that you are in. for on-prem setups resources are only available at a specific geographic location where servers are.8 ways public clouds are better than on-prem
- Fault tolerance-through regions and zones applications on the public cloud are safe from downtimes because where a zone fails or there is a regional outage service can rely on the other active region. It’s advisable for production workloads one employs the rule of three, where data is replicated across three zones to ensure that in the instance that one zone fails the production workload will run at over 50% supported by the two remaining zones. In on-premises setups, such guarantees are almost nonexistent because of the cost implication of buying servers and building extra data centres.
- Scalability-scalability is the ability to vertically/horizontally scale a workload in no time. on the public cloud, it’s easier to scale based on customer demand. Global resources like GCP cloud spanner horizontally scale database workloads across GCP regions ensuring low latency and high availability of services.in an on-premises setup, you will need to add servers to cover the periods where demand will be high which is very costly as a capital expense(CapEx).
- Global reach-public clouds are in the race to add as many regions as possible to bring resources closer to customers. Recently google cloud announced they intend to open a new region in south Africa a first in the continent in response to Azure and AWS who already set up shop a while back. Through their regions and network PoPs public clouds guarantee that workloads can have the lowest latencies and prices an advantage to consumers.
- Agility-this is the ability to react quickly to workload demands. Public clouds through their various offerings like load balancers can quickly respond to workload demands with ease. During the recent 2022 Amazon prime day, 2022 Dynamo dB processed 105 million requests a second. With on-premises, responding to such events will be very costly, especially where you have unpredictable spikes and surges in demand.
- Elasticity-elasticity and scalability are almost one and the same thing however for a service to be able to be declared elastic it has to be able to auto-increase and decrease almost in an instance. AWS Elastic Kubernetes services has the ability to auto-increase/decrease pods based on demand.
- Customer latency capabilities-if you search the web frequently chances are that you have browsed a site that uses Cloudflare. Cloudflare among other services is famous for its CDN which reduces the time taken to fetch a file from a server by storing the file in a PoP nearest to the user. years ago with the advent of the web, it would take time for a user in Australia to fetch a website in new york but now with CDN it only takes seconds.
- Predictive cost consideration– with on-premises it is difficult to predict the cost of running workloads because of outside factors like power, rent, security etc… that have to be put into consideration. all the major clouds have calculators that enable a user to calculate the cost of running a workload let’s say a compute engine workload for a month in any region. in Azure, there is also the TCO calculator that calculates the cost differences of running a workload either on-premises compared on the cloud.
Bonus.
- Security of applications-a massive DDoS attack will shut down all premise workloads but have little or no effect on the public cloud, due to the massive scale of cloud infrastructure and their multi-layered approach to security, it’s easy for them to absorb such hits without the customers ever noticing it.AWS guard duty and GCP cloud armor solve such issues with ease.
- Faster time to market-for public clouds it’s easier for developers to go from concept to consumer in a fraction of the time as compared to on-prem. this is because they are focused on solving actual problems and not worrying about servers and networking issues all of which are abstracted on the public cloud. with services like GCP app engine, a developer can easily split traffic to deployment to test the effects of changes in code in real-time.
