I'm attempting to compare cloud computing (on EC2) against traditional hosting about the following grounds to find out whether these features present unique benefits on the planet of cloud computing versus classical hosting methods:
- Real-time monitoring
- Server virtualization
- Deployment automation
- High end computing
- On-demand elasticity
So far as I can tell, (1) monitoring is equally as easy both in areas (2) server virtualization can also be contained in both areas because of server farms which permit traditional hosts to strengthen assets when needed - not to mention exactly the same is applicable within the cloud (3) deployment could be equally automated both in areas because the same tools frequently does apply to both (4) in the region of high end computing you may have an extra boost in the cloud theoretically but I am not too sure - you spend for your boost be it the cloud or otherwise (5) elasticity is really the only benefit will be able to see of moving towards the cloud - assets could be pumped up in the flick of the switch.
So my real question is, is the only real advantage of cloud computing out of this list that provides a genuine benefit over traditional hosting or perhaps is my analysis problematic?
The primary difference this is actually the cost model. While it's correct you will gain all the same advantages of your list with both Cloud Computing and traditional hosting, you repay front for traditional hosting. You need to buy and gaze after your personal servers, while cloud computing enables you to definitely pay a flexible cost.
Because of this , cloud computing is really attractive for startup companies.
You don't only have elasticity, however, you have, theoretically a minimum of, a larger total quantity of assets available than you might have with any static hosting solution.
Also, an unwanted effect of elasticity is decreased electricity usage, which might be considered a factor for you personally.
The organization Sometimes for is on the point of move from self-hosting to some cloud provider (EC2). One factor I'm greatly searching toward isn't needing to be worried about controlling hardware. I don't have to be worried about lead time for ordering parts. The requirement to have spares on-hands to pay for unpredicted hardware failures is finished. I don't have to be worried about UPS or any energy. We are really not large enough to cool down the to become a concern... however we won't need to bother about that either.
Depending by yourself datacenter costs, a cloud computing platform could be less expensive, as you do not need anybody to handle physical products. Cloud services can offer bulk computing assets at likely less cost than you are able to provide should you bought the machines and hooked them up yourself.
Presuming your "traditional hosting" involves just one server, there's an extremely real help to high-performance computing in cloud / power grid conditions. Particularly, virtually limitless performance, since you could have n cores working simultaneously, whereas having a single server, you're restricted to the most server capacity.
To place it more clearly, when the most effective computer on the planet is really a 1000 - core system with 20 terabytes of RAM, then this is the most energy you might have on the located server. However, a cloud composed of 100 of those machines could do 100x the job in almost the equivalent time.
Furthermore, it's generally less costly (financially) to distribute work across multiple more compact machines than to obtain one effective system able to do exactly the same work.
And when you want to discuss disaster recovery....clouds could be geographically distributed, meaning if your tornado rips your computer data center from the ground, plucks the server into little shards of metal and plastic, and embeds them in telephone rods...you have a small dip inside your performance since your other 99 servers continue to be operating.