Best practices for building large-scale systems on Azure platform based on open source software

Objective

Microsoft Azure is the only solution to Microsoft's public cloud. With this platform, users can deploy and publish their applications in a variety of ways. This is an open platform, in addition to Windows Server and SQL Server support, Microsoft Azure also supports a large number of mainstream open source software and frameworks, including Linux, Tomcat, Java and so on. How do you migrate an Open-source system to the Microsoft Azure platform and give full play to the benefits of the cloud in terms of elasticity, agility, etc.? This article and the reader share some best practical experience, and at the end of the paper to China Network TV system migration as a reference case.

Before you read this article, you want to know something about Microsoft Azure. Readers can log in to http://www.windowsazure.cn to get information about the platform. The following concepts are mentioned in this article:

-Infrastructure as services (Iaas:infrastructure as a service): Consumers through the Internet access to computer infrastructure classes of resources, including computing resources (virtual machine), storage resources and network resources.

-Platform as service (Paas:platform as a service): Cloud platform services providers to the resources on the basis of the IaaS to do further encapsulation, the IAAS layer to the user transparent. With the help of development tools and a series of development languages, consumers build their own applications on cloud platforms without the need to manage and maintain virtual machines, storage, networks and other infrastructures.

Using Linux virtual servers in Azure:

The Azure platform currently supports three types of Linux operating systems: UBUNTU, CentOS and SuSE. Their versions are:

-Ubuntu Server 12.04 LTS

-Ubuntu Server 12.10

-Ubuntu Server 13.10

-Ubuntu Server 14.04 LTS

-Openlogic

-SUSE Linux Enterprise Server SP3

Using Linux virtual machines on the Azure platform has the following best practices to help users get a better experience.

Configure swap partitions for Linux (swap)

There are two important reasons to keep swap partitions in Linux:

When physical memory is not sufficient to support the operation of systems and applications (processes), this swap space can be used as a temporary storage of low usage memory paging, freeing up memory to be used by badly needed applications (processes).

Even if your machine has enough physical memory, there are programs that can give away the physical memory space by transferring memory paging content that is rarely used when they are initialized to the swap space. For applications (processes) that have a memory leak probability, swap space is more important because no one wants to see the system crash due to insufficient physical memory.

On the Azure platform, when a user creates a Linux server through a mirrored library, no swap partition is assigned by default. To run the free command, we can see that the corresponding number for swap is 0.

See more highlights of this column: http://www.bianceng.cnhttp://www.bianceng.cn/Servers/cloud-computing/