Cloud computing update

After all, the maiden article in this column, back in 2011, was devoted to cloud computing. Cloud computing could save your business tons of money and improve your turnaround time if you are in the business of rolling out software applications. The “cloud” in cloud computing of course refers to the Internet. So, essentially, cloud […]

After all, the maiden article in this column, back in 2011, was devoted to cloud computing. Cloud computing could save your business tons of money and improve your turnaround time if you are in the business of rolling out software applications. The “cloud” in cloud computing of course refers to the Internet. So, essentially, cloud computing is “computing on the internet.” That is, from your desktop, laptop, or smartphone, you could log into some “far-away” website and start “computing,” which means running some applications or developing yours in other people’s infrastructure located on the web.
The cloud infrastructure is essentially that of a data center, with several utilization modes. Depending on the scale and the degree to which the cloud allows other people access, we could have private clouds, public clouds, and hybrid clouds (private plus public). The first mode of utilization is “Infrastructure as a Service (IaaS),” whereby you upload your application to the cloud site and run it there for a fee, using the hardware (servers, storage capacity, network bandwidth, etc.) at the site. The second mode of usage, “Platform as a Service (PaaS),” allows you to log into the cloud site and use software tools in the site to build your own applications. This obviously enhances quick roll-out of your software. In the third utilization mode, “Software as a Service (SaaS),” you access the cloud site in order to use available software (developed by others) on the site. The cloud supplier, not the user, is responsible for the hardware that delivers this service, and also for the creation, updating, and maintenance of the software. In all cases, you pay only for what you use.
In my experience, not everyone who says they operate a cloud facility actually does so, in the strict sense of the technology. To recognize a cloud deployment, look out for the following characteristics. Clouds are, of course, web-based, and usually involve a large number of servers (machines), which are built from proven commodity parts. As an example, the word is that Google’s cloud has at least 800,000, and perhaps one million CPUs, dispersed over at least twelve geographical locations of the world. The microprocessors in cloud servers are typically of the x86 (generic PC) variety and each server is virtualized to hold anywhere between 8 to 256 users or more, through the use of multiple cores per CPU and a couple of CPUs per machine. Also, cloud computing involves a business model that is characterized by elastic resources that are available on demand to whomever needs them.
Many institutions, especially in the United States of America, such as Cornell University, University of Illinois at Urbana-Champaign, University of California in San Diego, and University of Texas at Austin, have built out massive supercomputers from a cluster of small machines. However, those supercomputers would not qualify as cloud installations.
A few companies in Nigeria have tinkered with cloud deployments, although this has mostly been limited to placing applications in cloud installations abroad (mostly in the USA) and having customers (usually in Nigeria) log into those clouds to use the software. AppZone Limited in Lagos previously housed its application software with the Go-Grid cloud (USA), where customers went to use the software. TTC Technologies (Nigeria) Limited, in collaboration with the Obafemi Awolowo University (OAU) in Ile-Ife, is the only Nigerian company that I know of that is actually building (private) clouds, by carrying out the requisite cluster computing tasks.
Should your business or institution go into the cloud? Numerous advantages abound for you with the cloud. The capital investments you need in order to replace hardware and infrastructure every three to six years disappear, as do the operating costs, maintenance costs, software licenses, power, and so on. In its place, you pay for only what you use, and you are not stuck with redundant facilities outside of your peak demand. Because of elasticity, the amount of resources available to you scales with your need, seemingly without a limit on the resources you can ask for and get. The downside to cloud computing is that information security could be an issue, and, depending on your business, you could opt for private clouds.
The cloud business is still growing, and previously skeptical companies, notably IBM and Oracle, have now fully joined in the cloud computing bandwagon. In fact, both companies now offer a broad selection of enterprise-grade cloud solutions that encompass all the three utilization modes. Amazon is, by a large margin, presently the biggest player, with revenue last year of more than $3 billion, an increase of over 85% relative to that for a year earlier. In the past year, Google has revamped its offerings, an event that has motivated customers to jump to Google from both Amazon and Microsoft.
Another new development in cloud computing is the rather brutal price war that is going on between the major cloud providers: Amazon, Microsoft and Google. Basically, the economy of scale brings the cost of computing on the cloud to extremely low rates. A few years ago, Amazon was charging approximately N13.6 an hour per server to run on a Linux server and N19.2 to run on a Windows server, while Microsoft charged N20 per hour per server. Shira Ovide pointed out a few days ago in the online Connect magazine that, within days last month, each company had cut prices on various cloud services by up to 85%.
The bottom line of this article is that more providers and users are getting into cloud computing, and price wars are driving affordability way up.