PaaS is popular in certain segments. As the list of the use case of IaaS and SaaS is too long, it does appear that PaaS is a failed cloud computing model. One of the great benefit of cloud computing is agility. PaaS move agility to another level but that is not what every segment wants. Theoretically, abstraction of the underlying infrastructure can help to focus more on business requirements. But in real life, PaaS faces certain challenges which drive away from both the general consumers.
The Confusing Marketing of PaaS
Since the year 2010, we have written many guides on PaaS. Many of the readers were interested in RedHat OpenShift because of the free quota. Heroku cloud already had a good number of users. In these 10 years, PaaS has been super-complicated. There are public, private, hybrid, managed service, on-premise, hosted PaaS. No blogs can write stable guides which will work for a long period. When peoples think about a product, they check reviews and if required, follow the how-to guides. Landing to Openhift’s official webpage :
Any level of the consumer will get confused. A newbie IT manager after reading “…Kubernetes container platform with full-stack automated operations to manage hybrid cloud, multi-cloud, and edge deployments…” will start to perform research on each keyword. And then the phrase “to improve developer productivity and promote innovation” will drive away most of the seasoned IT managers. It is not just about RedHat or IBM. They face a lot of jargon with Azure too.
This problem appears because PaaS has no universal standard. Every vendor wants to deliver a message about their special features.
Targeting by Trial and Error
PaaS is only partially comparable with shared hosting. There is nothing on this earth that is like PaaS. If you need PHP for hosting applications, it is easy to use shared hosting. For lack of guaranteed SLA and being an immature model, the target audience goes wrong. It is often said, “PaaS is great for the developers”. In real life, many developers try PaaS for the free tier. When they work as an advisor of a company, they suggest preferring IaaS and SaaS. PaaS got almost no clients from those who are looking for a solution for hosting websites.
The companies know the challenges, they targeted the mobile app developers. But, today the cost of cloud servers is lower, getting a virtual dedicated resource is the better option to them because the software is essentially free and some tweak at the backend forces to perform better. Most of the enterprises need Java and .Net and there is nobody good at supporting both. So the companies now re-targeted data analytics and IoT segment. Possibly, these are segments where PaaS can become a standard.
The Learning Curve
It is not just easy to get used to one PaaS and shift to another. While it is true that today OpenShift is being hosted by different vendors, but in reality, you can not easily move from IBM to AWS.
The Cost Factor
Unfortunately, the running cost of PaaS is higher. So in the end, per GB pricing of PaaS becomes similar to per GB pricing of IaaS. At $10/month, you get a fully-featured cloud server (a part of IaaS), who will be willing to pay $10 for using the free PHP instance of PaaS? Very few use PHP, Python or PaaS spending thousand of dollars per month. The theoretical idea behind PaaS is great but creating the backend for a simple application is easy. Again, the users of data analytics and IoT segment find PaaS useful but most enterprises look for private or hybrid solutions. PaaS is not always reliable to them.
When PaaS took birth, the solutions were from Google App Engine, Force.com, Heroku, Engine Yard, etc. It was a different era when IaaS, SaaS and different software development methodologies were not like today. The public cloud was not easy to set up by the startups and SMBs. Today there is a growth of newer cloud computing models, such as FaaS. So, the classic PaaS which could be used for PHP hosting has no bright future. PaaS will remain and grow in new forms.