A few days back I posted an article on Top 10 Cloud Computing Service Providers. Many readers asked what is Cloud Computing? Why is there so much buzz about Cloud Computing in the tech-industry? Why is its definition so fuzzy?
I believe we understand better with examples and practical scenarios. These scenarios may help you in understanding what is Cloud Computing and how it may help your business. Also, please note, this is an introduction and cloud computing is indeed much more than what is discussed here in this article series.
[Updated on Nov 29, 2011]
August 2008, United States of America
At the end of August [2008], as Hurricane Gustav threatened the coast of Texas, the Obama campaign called the Red Cross to say it would be routing donations to it via the Red Cross home page. Get your servers ready—our guys can be pretty nuts, Team Obama said. Sure, sure, whatever, the Red Cross responded. We’ve been through 9/11, Katrina, we can handle it. The surge of Obama dollars crashed the Red Cross website in less than 15 minutes.
—Newsweek
Feb. 21, 2011. During ICC Cricket World Cup, India
Furious cricket fans slammed organizers of the World Cup on Monday as the official ticketing website crashed amid a scramble for 1,000 tickets available for the final…
---The Economic Times [Source]
The official ICC partner for online ticket sell, Kyazoonga, posted the following message on its Facebook Fan Page and other social networks.
We are facing absolutely unprecedented amounts of traffic from all over the world, with hundreds of millions of people hitting at once. Some of you may have trouble accessing the site. It seems that cricket fever has surpassed all anticipations and expectations. Please bear with us as our global network team works on bringing you the tickets you all have been waiting for.
June 2009, China
A Chinese website set up so people can inform on corrupt officials has been inundated with so many visitors that it crashed shortly after launching.
---BBC
Perhaps this example made you smile, or perhaps you are equally worried as well, regarding the level of corruption! No need to worry, because this also signifies a huge number of proactive citizens willing to report it. Look on the bright side.
Everyday 8:00AM to 9:00AM, Since ? Till Date, Online Ticket Booking - IRCTC, India
…The bookings & enquiry requests are 6-7 times higher during the peak hours than the rest of the day. So while the existing infrastructure is well equipped to handle the daylong traffic, it is the peak hour traffic that clogs servers…
---What IRCTC can Learn from redBus Cloud Implementation?
May 2009, India
The world’s largest democracy, 1 billion+ population, goes to the jumbo general election. The election commission unveiled a brand new website for providing real time results of the mega-poll. It showed off arrangements which indicated it was well prepared to handle 80.64 billion hits in 8 hours (2,800 hits/second). Obviously a decent number by any standard.
Guess what happened?
On the election result day media reported:
300,000 hits/second make Election Commission website crash.
This per second hit rate means 8.64 trillion hits in 8 hours. Is it less than Google’s hit/second? Take a guess.
Similar server crashes were reported across India during the online CAT examination conducted by prestigious IIMs (Indian Institute of Management) in November/December 2009, though they were smartly attributed to a virus and not to the number of hits.
October 2009, France
A website launched by French first lady Carla Bruni has crashed on its first day - overwhelmed by the number of users trying to access it at once.
---BBC
Online mob. Wish I was even half as popular as her!
What inference can we draw from above server crashes/failures?
These examples of server crashes are but a few among the hundreds happening almost everywhere in the world on a regular basis. What does this indicate? Just one conclusion and that is:
The situation can only get worse. Currently, only 25% of world population, i.e., approximately 1.75 billion people have Internet access. Compared to Television and other mass media, it’s still considered an elite medium of communication. If this meager Internet penetration has lead to a significant number of crashes, what will happen when the Internet becomes a mass medium? India and China currently have approximately 15% or less Internet penetration. Even a moderate increase in Internet penetration and usage, say to 50%, will add more than a billion Internet users! Definitely, too many clicks to handle. Going by statistics and backed up by almost double digit GDP growths, they are soon going to add to this number.
What’s the solution to prevent the above scenarios?
Most of you will answer:
● Add more servers to balance the load (load-balancing)
But will adding more servers solve the problem, or compound the problems of companies/organizations that are already on a tight budget?
Moreover, how many servers do you think will be enough to handle that kind of traffic? Take a guess. 100? 1000? …
Before you guess, let me provide some facts to help you make a wise guess.
● Facebook uses 30,000 servers! (as of October, 2009) and it’s adding capacity on a daily basis. Sounds weird.
● An unofficial estimate predicts the number of Google servers to be an incredible 1 million in its world wide data centers!
The websites discussed above are not as big as Facebook or Google but their spike in traffic on that particular day may had beaten these giants! Given these kind of stats, in order to purchase the necessary number of servers, the organizations/entities involved will fork out enough money to make them eligible to file bankruptcy protection.
So, where’s the catch? Even if they are financially sound enough to add huge numbers of servers, just remember the following from our above scenarios/case study:
● The Red Cross will get this exponential surge in the number of hits once in a decade or maybe we don’t know when (i.e., only when there is a natural calamity of larger scale)
● The Indian election commission website will attract visitors only when there’s an election. i.e., ideally once in 5 yrs (forget regional elections, they don’t attract much traffic).
● The Chinese corruption website traffic would have gradually reduced to a normal level in a month or so.
● Carla Bruni’s fans would have mobbed her website only for a week or so.
The above description clearly shows the following trend:
Most of the traffic spikes are predictable and can be planned for. Even the Red Cross traffic surge was predictable, but you’ve got less time to react and plan for it. So adding thousands of servers to handle a few days or seasonal spikes in traffic is a humongous waste of resources.
So, what do you think is the best solution? Do share your views.
Stay tuned to Techno-Pulse for the 2nd part of this article on Cloud Computing Introduction – 2, we’ll analyze further and find the best solution.
Related Cloud Articles
- Memorable Cloud Computing Quotes
- Infrastructure as a Service – IaaS: Cloud Computing
- Cloud Computing Platform Introduction – PaaS
- SaaS Introduction with Example – Cloud Service
- India Based Cloud Computing Service Providers
- Cloud Service Models - SaaS PaaS IaaS - Which One is for You?
- Cloud Deployment Models – Private, Community, Public, Hybrid with Examples
