Database Reference
In-Depth Information
Dell Optiplex GX620, equipped with Intel Pentium D CPU 3.00 GHz, 2 GB
memory, and 10/100/1000 Base-T Ethernet.
Both Amazon EC2 and Microsoft Windows Azure instances use the default type,
small instance with single core.
Qualitative Experience of Development Utilities
In Amazon EC2, an administration role will be granted to developers when a virtual
machine instance is created. This allows the developers to install whatever they
want in the instance. In other words, there is no restriction on selecting development
environments for Amazon EC2. But on the other hand, being able to select different
work needs to be done, such as uploading and installing the required runtime
environments for the application.
The key highlights of the Microsoft Windows Azure platform are its heavily
equipped frameworks and environments. Almost all existing Microsoft web devel-
opment frameworks and runtime environments are supported in Microsoft Windows
Azure. As a result of this, developers can simply focus on the business logic
implementation with C# or PHP. But the key downside is that they have to stick
with Microsoft development environments, Microsoft Visual Studio.
In contrast to Microsoft Windows Azure which offers fully functioned frame-
works, and Amazon EC2 which provides highly configurable environment, Google
App Engine re-implements programming languages to suit the different devel-
opment approaches. Google has currently enabled Python and JVM-supported
languages on its cloud platform where developers are free to choose frameworks
based on Python and JVM-supported languages to improve their productivity. But,
in practice, there are some limitations on the Google App Engine which restrict the
range of choices, such as no multiple threads, no local I/O access, and 30 s timeout
a request handler. Additionally, Google also offers other Google APIs to integrate
Google App Engine with other Google services.
Quantitative Results of Test Sets
High Stress Round-Trip
Figures
4.4
-
4.6
indicate the cumulative distribution function of response time under
varying amount of concurrent stress requests, which range from 300, 900, 1,500,
2,100, 2,700, up to 3,300 requests respectively.
The observation of three cumulative distribution functions confirms that the
larger the requests, the longer the response time will be. But the incremental step
of response time varies from one group of requests to another, depending on the
cloud hosting servers. For 80 % of cumulative distribution functions, the response
