Gartner Analyst Nick Jones posted an interesting article in his blog with the title The cloud computing fantasy. Contrary to many analysts, he is a cloud skeptic. The argument he puts forward is my favorite one against the cloud computing hype, although I wouldn't go so far as to say that cloud computing is a fantasy.
- Pip install Boto3 - Thu, Mar 24 2022
- Install Boto3 (AWS SDK for Python) in Visual Studio Code (VS Code) on Windows - Wed, Feb 23 2022
- Automatically mount an NVMe EBS volume in an EC2 Linux instance using fstab - Mon, Feb 21 2022
On the contrary, cloud computing has been a reality since the early days of computing. Not so long ago computation took place only in huge datacenters. Cloud computing never really died out. Basically, Nick Jones' blog post names one of the reasons why cloud computing got a little outmoded: storage capacity for home users (and therefore also for businesses) became cheaper than network connections.
I think his prediction that by 2012 we will have cheap terabyte flash drives is realistic. I just bought a new eSATA disk drive with 1.5TB. It cost me only 200 Euro. All in all my laptop now has 3TB storage capacity. There was a time when this would have been an unbelievably big storage capacity for a whole datacenter. Moreover, if you rent this storage capacity nowadays at a cloud provider you would pay significantly more. Therefore, it is not just the costs for the network connections to access the data, but also the fact that the prices of cloud providers for storage capacity are much higher.
I'd like to add that this doesn't only apply for storage, but also for computational power (hardware and software). Ever since the PC invaded our offices and homes there has been a race going on between on-premise computing and cloud computing with regard to prices and capacity for computation, storage and data access. For decades, on-premise computing has been the unbeatable champion in all disciplines.
However, it seems to me that cloud computing is catching up lately. There are two reasons: First of all the Internet became a mass phenomenon. This pushed network speeds and lowered the prices for remote data access tremendously. Second, PC technology is replacing expensive mainframe technology in the datacenters. In this way, cloud computing profits from the low prices of the mass market for the PC technology.
Both phenomena are not new, though. Prices for cloud computing are still higher than for on-premise computing. Thus, it remains to be seen which of the two opponents will be the winner in the coming decades.
Of course, the race is not only about the mere costs for computation, storage and network. It is also about manpower. The management costs often outweigh the other three factors. It is the main argument of cloud supporters that management costs will be lower when all the computation takes place in datacenters. However, it escaped their notice that in the past on-premise computation was also the champion in this category. Operating a huge datacenter costs a lot of money because you need many high-paid specialists. This is probably the main reason why many cloud providers are expensive nowadays.
However, it is quite likely that the prices for datacenter specialists will fall soon. The more datacenters there are the more specialists will flood the market. Furthermore, the datacenter management tools will improve fast now. Therefore, I believe that cloud computing could catch up in the long run. But we should not forget that on-premise management is improving fast too.
I think that there is a general misunderstanding with respect to cloud management. Many believe that if computation takes place at one location it must be cheaper than if it is spread over many different locations. While this may be so, this factor is often overestimated.
On the one hand the improved network speed made the renaissance of cloud computing possible in the first place. But it also improved the management of on-premise computers. It is now possible to easily manage PCs that are on the opposite side of the globe. Centralized management does not necessarily entail that all the computation has to take place at one physical location, the datacenter.
You could say that the Internet made the virtualization of the physical location of computation possible. It simply doesn't matter anymore where computation happens. Actually, this was the main motivation of the scientist who invented the Internet. The Pentagon wanted to decentralize computation for reasons of security. Therefore, improved network speeds will also work against cloud computing because it reduces the costs of centralized on-premise computing. From a system administrator's point of view it will not make a difference if a desktop system is running on a physical PC at an employee's home or in a virtual environment in the datacenter.
I only covered a few factors that influence the race between cloud and on-premise computing in this post. Other considerations I have made can be found in Is cloud computing a threat for Windows?, Is cloud computing secure?, Will cloud computing reduce software prices?, and Will cloud computing render IT administrators jobless?
Subscribe to 4sysops newsletter!
In my opinion it is impossible to predict the outcome of this ever ongoing race in the near future. It is like with horse races. You might have good reasons to bet on the celebrated newcomer or on the old champion. (I will leave to you to decide which is the newcomer.) Betting in horse races is mostly gambling. The same applies to predictions about the future of computing.