Comparison of cloud and grid computing based on performance [closed]

Question

Closed. This question is opinion-based. It is not currently accepting answers.

---

Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.

Closed 7 years ago.

Improve this question

how cloud computing and grid computing differs on the basis of performance. say for example .. i have a local grid and a cloud service .. which one will be better to use if performance is my only criteria!! i have gone through many differences on cloud and grid computing but found nothing significant over performance issue.

P.S :- i am bit new to cloud computing

Answer 1

Generally speaking, cloud computing will give you slower machines, and all your data will have to go over the internet. However you can buy machines by the hour, so that for bursts of compute power you can probably afford to get a lot of machines.

If you set up your own local grid, you'll be able to decide the spec of machines that you get, and set them up how you like. However it'll be expensive to buy a ton of them if you're not using them all the time.

If you want to do the cost analysis, you can take a look at our cost calculator here (https://secure.slicify.com/Calculator.aspx) to get a rough idea how much on-demand compute power will cost from a variety of providers (full disclosure - this is a page on my site).

Answer 2

Impossible to say. There is no winner. Your 'performance' is based on how fast the physical machines are (CPU/RAM/IO), your work load and the inefficiencies of the distributed system. Which type of distributed system is best for you depends on your work load.

Sorry, but if you formulate a better question, you might get a better answer.

Answer 3

Well, with Cloud, you need to ask how much data you need transfer via link between machine hosting the calculation and the data source? And how fast. You mentioned you didn't detect significant differences? You probably wired very small amount of data.

Answer 4

I don't think that's ... the correct question to ask. For example, Google Compute Engine charges based on what hardware you're choosing.

In general, your code can run as fast in Cloud (does not equal "Grid") as in local LAN. You need figure out where (if there exists) the bottleneck is. All "Grid" implementations farms out calculation to "Nodes". But bottleneck on database tier.

Hadoop is one which get around this by "HDFS" (Hadoop Distributed File System).

Obvious alternative is, for example, database clustering.

Anyway, in general, there's no reason why your code run slower in Cloud (Please don't tell me link between database and your grid is slow however!)