Windows Azure HandBook (2) Azure China-Alibaba Cloud Developer Community

Catalogue of Windows Azure Platform series articles

 

  for traditional self-built Data centers, enterprises need to manage the underlying Network,Storage,Servers,Virtualization, middle-layer OS,Middleware,Runtime, and top-layer Application, and Data. This is like self-built houses in rural areas.

Public cloud platforms are generally divided into three types: IaaS, PaaS, and SaaS.

  Microsoft Azure platform belongs to IaaS and PaaS.

 

 1. IaaS

for users, the underlying Network, Storage, Server, Virtualization Microsoft Azure are all ready. You only need to deploy applications to the Azure IaaS. You only need to manage the middle-tier OS,Middleware,Runtime, and top-layer Application,Data.

Azure supports the following operating systems:

-Windows : Server 2008 R2, Server 2012, Server 2012 R2

-SQL Server: SQL Server 2008 R2, SQL Server 2012 SP1, SQL Server 2014 RTM (Web, Standard, Enterprise)

-Linux:

Ubuntu (12.04 LTS, 12.10, 13.10, 14.04 LTS)

CentOS (6.5)

SUSE (SUSE Linux Enterprise Server 11 SP3)

for IaaS, the management and maintenance of the above parts of the OS need to be managed by users. This disadvantage is that the for example user now use Windows Server 2012 R2, future Microsoft release. Windows Server 2012 R2 SP1, REQUIRE user upgrade manually.

  

2. PaaS

in PaaS, OS,Middleware, and Runtime are all managed by Microsoft Azure. PaaS emphasizes providing platform capabilities without the need to care about OS,Middleware, and Runtime. That is, the customer only needs to tell the Microsoft Azure to provide the ability to run Application, whether it is. NET, Java, PHP, etc. Enterprises can easily deploy applications to Microsoft Azure. Microsoft provides Azure SDK development platforms:

-.NET

-Java

-PHP

-Node.js

-Python

-Ruby

in terms of migration difficulty, it is more difficult to migrate data from IDC to Azure PaaS than IaaS, because PaaS requires some code modifications.

My personal suggestion is to allow users to migrate local applications to Azure's IaaS platform to enjoy the convenience and reliability brought by cloud computing. Then, on the premise that IaaS runs smoothly, migrate some applications to the PaaS platform.

 

3. SaaS

Microsoft Azure is not a SaaS platform.

SaaS is an out-of-the-box software solution for users. End users do not need to know which cloud computing provider is deployed behind the software solution and whether the platform behind it uses IaaS or PaaS Services.

Microsoft's Office 365 is a typical SaaS service. Solutions like Salesforce.com also provide SaaS services.

For those ISV vendors, the traditional way to deliver software requires the following steps:

-First deploy hardware services on the customer site, including installing servers, networks, firewalls, and power supplies.

-Install software, including operating system, database software and runtime software

-Install application software

as you can see, such a software delivery process actually includes the delivery of hardware and software. For ISV vendors, they need to maintain not only the availability of software, but also the underlying hardware servers.

After ISV manufacturers adopt Microsoft's Microsoft Azure public cloud services

-IaaS: delivers the underlying Network, Storage, Server, and Virtualization to Microsoft Azure.

-PaaS: delivers the underlying Network, Storage, Server, Virtualization, and middle-tier OS,Middleware, and Runtime to Microsoft Azure.

ISV vendors only need to focus on software development and upgrade. The O & M and other work of other hardware servers will be handed over to Microsoft Azure.

By transforming and upgrading applications in this way, those traditional ISV vendors can use Microsoft's public cloud services to provide SaaS services to their customers.

  

  

 

Services provided by Microsoft Azure

  note: this topic only describes the services provided by Azure China. Azure platform is an updated platform with a new feature release every month. The author tries his best to update the document, but this article does not represent the final service content provided by Azure platform.

According to the business value of the Microsoft Azure and typical business scenarios, the service components of the Microsoft Azure are classified as follows:

  

note: the above content will increase as Microsoft Azure platform changes. For more information about all services provided by Azure, see www.windowsazure.cn.

Based on these service components, users can flexibly build their own IT architectures and application environments.

  

1. Storage services

  Azure storage is a file storage service in the cloud. You can store local binary files, such as files, pictures, and photos, in the cloud.

In a traditional IDC data center, storage is a machine name, or stored under a disk of a server, or a network location of a storage.

The Azure storage service is actually an http / https network path that allows permission control. Azure storage does not depend on any IP address or network path.

  

The storage service itself supports a 99.9% SLA, which provides three types of high availability :

  1) triple redundancy (Local Redundant Storage and LRS) of the Local data center. For example, customers can choose to implement triple redundancy for storage services in the same data center, such as triple redundancy for data centers in Shanghai. Any file stored in Shanghai storage service has a primary backup and a sub-backup.

  

For example, if a customer uploads a 10GB movie, the Azure storage service actually stores 30GB in the same data center. However, Azure charges only 10GB of the actual Upload fee.

For LRS, the triple redundancy of transactions in the same data center is executed synchronously.

 

2) triple redundancy across data centers (Geo Redundant Storage, GRS)

  careful users will find that Microsoft's data centers are built in pairs at home and abroad, such as Beijing Data Center and Shanghai data center. This is because Microsoft has fully considered the capability of remote redundancy. There is a leased line between the data centers in Beijing and Shanghai. This leased line is dedicated to data synchronization before the intranet data center.

  

For example, you have created a storage account in the Shanghai data center and enabled cross-data center synchronization. When a user uploads a 10GB movie to Shanghai data center, the movie file not only has triple redundancy in Shanghai data center, but also has triple redundancy in Beijing Data Center (auxiliary location), A total of six redundant files. For example, even if the Shanghai data center was completely destroyed due to earthquakes, wars and floods, users' data were still safely stored in Beijing's data center, and the files were truly safe.

For a data center in the primary location, transactions are asynchronously sent from the primary location to the secondary location.

The following table shows information about the current primary and secondary locations:

  

   

  

GRS the RPO and RTO what is it ?

Restore Point target (RPO) : in GRS and RA-GRS, the storage service asynchronously replicates data from the primary location to the secondary location across regions. When a major regional disaster requires failover, the latest incremental changes that have not been replicated across regions may be lost. The number of minutes of potential data loss is called RPO (that is, the point in time to which data can be recovered). Our RPO usually takes no more than 15 minutes, but there is no SLA concerning how long cross-region redundant replication takes.

  

Recovery time target (RTO) : Another metric that needs to be understood is RTO. This refers to the time required to perform failover and bring the storage account back online after failover. The failover time includes:

(A) investigate and determine whether data can be recovered in major locations or whether failover should be conducted within the time required

(B) time for account failover by changing DNS entries

we attach great importance to protecting your data, so if there is any possibility to restore data, we will suspend failover and try our best to restore data in major locations. In the future, we plan to provide an API that enables customers to trigger failover at the account level so that they can control RTO on their own, but it is not yet available.

  

3) Read access region redundancy (Read Access-Geo Redundant Storage, RA-GRS)

  simply put, if you create a storage account in the Shanghai Data Center (the main location) and enable the RA-GRS, the transaction is asynchronously copied to the data center in Beijing. RA-GRS provides read-only access to data centers (secondary locations) in Beijing to achieve higher read availability for storage accounts. In this way, you can specify whether the access to the Azure Storage points to the Shanghai Data Center (primary location or Beijing Data Center (secondary location) to improve the read availability. When this feature is enabled, you can use secondary locations to read data in major regions for higher availability. Select use. This feature requires cross-region redundant replication of storage accounts.

 

For example, assume that I have created a Azure Storage in the Shanghai Data Center (the main location), the Storage Name is leizhangstorage, and the Read Access-Geo Redundant Storage, RA-GRS is enabled.

(A) I can access the http://leizhangstorage.blob.core.chinacloudapi.cn of the main location through Azure Storage Account.

(B) http://leizhangstorage-secondary.blob.core.chinacloudapi.cn of secondary locations can then be accessed through Azure Storage Account

(C) in the event that the Shanghai Data Center (main location) cannot read data, data reading at the secondary location can be used to provide high availability.

References:

http://blogs.msdn.com/ B /windowsazurestorage/archive/2013/12/04/introducing-read-access-geo-replicated-storage-ra-grs-for-windows-azure-storage.aspx

http://blog.csdn.net/azurechina/article/details/22749157

 

Azure storage provides three types of storage services: Blob, Table, and Queue.

  

  1. Blob

Blob stores large binary objects, such as files, images, and documents in binary format.

Blob can be divided into two types:

(1)Block Blob

this type is suitable for storing binary files and supports resumable Upload. It can be a block unit of up to 4MB. A single file can be stored for up to 200GB without continuous storage, it may be stored in different storage servers. Optimized for file upload and download. Block can be created in two ways. You can call the Block Blobs operation to upload a PutBlob that does not exceed 64MB. Block Blobs larger than 64MB must be uploaded in blocks, and the size of each Block cannot exceed 4MB.

(2)Page Blob

this type of storage optimizes random access. It is divided into a continuous area in the storage area to store data, which can be used to store VHD. A single file can store up to 1TB.

 

Blob service consists of Blob itself and its storage containers (Container). Containers can be regarded as local folders.

You can access Blob through REST API

http :// <accountname>.blob.core.chinacloudapi.cn/ <containername>/ <blobname>

accountname indicates the resource under the Azure storage account, which is globally unique. blob.core.chinacloudapi.cn indicates that azure china blob storage resources are fixed. containername indicates the name of the container. blobname indicates the name of the resource that I want to access. It can be considered as an mp3 file or a jpg file.

  

Example:

I saved it under the leizhangstorage account. The containername parameter is photo, and the blobname parameter is myphoto.jpg. The URL is as follows:

http://leizhangstorage.blob.core.chinacloudapi.cn/photo/myphoto.jpg

 

I saved it under the leizhangstorage account. The containername parameter is vhd, and the blobname parameter is myvm.vhd. The URL is as follows:

http://leizhangstorage.blob.core.chinacloudapi.cn/vhd/myvm.vhd

 

note the naming rules for Container.

-containername can only be a level -1 Directory. You cannot set the next level of containername under containername.

-Must start with a letter or digit, and the name can only contain letters, digits, and dash(-).

-Cannot start or end with dash(-). dash(-) cannot appear consecutively.

-All English characters must be in lowercase

-Length: 3 to 63

 

Blob naming rules

-Except for reserved characters of the url, other character combinations can be used

-The name must be 1 to 1024 characters in length.

-Avoid ending with dot(.) or forward slash(/). Otherwise, Blob Service misjudgment may occur.

 

2.Table

the Azure Storage Table here are non-relational data tables and cannot be confused with SQL Server tables. You can approximately consider Azure Storage Table as NoSQL.

  

Each row in the Azure Table is an Entity. The maximum capacity of a single Entity is 1MB. The maximum capacity of all records in the Azure Table Table is 200TB. Each Azure Table must have Partition Key and Row Key. A maximum of 255 attributes can be Azure Table. The value of the Partition Key can be used to set the physical location of the record.

If the Azure Table values of the two data records in the Partition Key are the same, the physical locations of the two data records are the same;

if the Partition Key are different, the two pieces of data may be stored on the same physical media or on different two physical media. As shown below:

  

Table is suitable for scenarios where log files are stored or non-relational databases are required.

 

  3.Queue

Queue is a First-Come, First-Serve (First-Serve) storage service, or FIFO (First-in, First-out). The queue can be a string or binary data of up to 64KB. There is a very important concept in Azure PaaS called Web Role/Worker Role. Queue serves as an important bridge for Web Role/Worker Role communication. The Azure PaaS/Web Role of the Worker Role platform will be introduced in detail in the next chapter.

 

Storage Service Performance Metrics

  one aspect that users need to pay attention to before choosing to use Azure storage service is its performance indicators, that is, whether the storage service can meet users' daily use needs, whether the peak access is met. The following table lists the maximum performance metrics of a storage account.

(1) the maximum data storage capacity of a storage account is 100TB.

(2) the maximum Transaction of a storage account is to process 5000 times per second (including Blob,Queue, and Table)

(3) the maximum bandwidth of a storage account is to transmit 3GB of data per second

the preceding performance metrics are the maximum performance metrics provided by Microsoft Azure storage services. Therefore, the performance metrics that users can obtain during application running are lower than these metrics.

  

Windows Azure use data partitioning to improve data access performance and auto scaling. Therefore, users should use data partitioning to distribute data to multiple servers to achieve the highest possible data access performance. The maximum performance metrics for each data division are as follows:

(1) for a single Blob, 60MB of data can be transmitted per second

(2) for a single Table Partition Key, 500 entity records can be processed per second

note that this is only a performance indicator of a single Table Partition Key, not a single Azure Table. Therefore, if a Azure Table can have a good Partition design, the performance can be improved.

(3) for message queue, 500 messages can be processed per second

 

if the application has reached the maximum performance of a single storage account, you can consider the following policies to further improve performance.

(1) for Blob, CDN can be considered to cache frequently accessed data to the nearest Edge Server to users to reduce data transmission time.

(2) for tables, we can consider designing Partition Key and try to distribute and store data in multiple places.

(3) for queues, you can consider centralizing multiple messages into one message or using multiple message queues.

 

Storage service security

  Windows Azure storage service is assigned two accesskey pairs: the primary accesskey and the secondary accesskey.

The accesskey is a 256-byte string. You can use the accesskey to verify each data access request to the Windows Azure storage service.

The workflow is as follows:

(1) the user first constructs a data access request, which can read Blob, write Entity to Table, or use Queue

(2) users use accesskeys to digitally sign data access requests through encryption algorithms

(3) the digital signature is included in the Header of the data access request (HTML Header), and then the data access request is sent to Windows Azure storage service.

Note that when you apply for a storage account, the system generates two accesskey pairs: the primary accesskey pair and the secondary accesskey pair. Both accesskey pairs have the same function to access data. There are two reasons to facilitate user switching of accesskey pairs. For example, if you use a master accesskey, you can disable the primary accesskey in the shortest possible time (regenerate a new primary accesskey) and switch to use the secondary accesskey. You do not need to modify the code, but only need to change the service configuration file, and update the service configuration file without downtime. In addition, an effective method is that the application code can first use the primary accesskey to access data. If the verification fails, the code automatically uses the secondary accesskey, in this way, users do not have downtime when switching access keys.

To further improve the flexibility of Blob data Access control, Windows Azure data storage service also provides another data Access control method called Shared Access Signature. With Shared Access Signature, the administrator can generate a temporary data Access request for other users. The temporary data Access request contains a valid digital Signature, so it can pass data verification. The temporary data access request also contains information such as what data and Operation permissions the user can access, and the temporary data access request is valid only for a specified period of time. In this way, the administrator can allow other users to access data without revealing the accesskey pair.

  

Note that when you apply for a storage account, the system generates two accesskey pairs: the primary accesskey pair and the secondary accesskey pair. Both accesskey pairs have the same function to access data. There are two reasons to facilitate user switching of accesskey pairs. For example, if you use a master accesskey, you can disable the primary accesskey in the shortest possible time (regenerate a new primary accesskey) and switch to use the secondary accesskey. You do not need to modify the code, but only need to change the service configuration file, and update the service configuration file without downtime. In addition, an effective method is that the application code can first use the primary accesskey to access data. If the verification fails, the code automatically uses the secondary accesskey, in this way, users do not have downtime when switching access keys. To further improve the flexibility of Blob data Access control, Windows Azure data storage service also provides another data Access control method called Shared Access Signature. With Shared Access Signature, the administrator can generate a temporary data Access request for other users. The temporary data Access request contains a valid digital Signature, so it can pass data verification. The temporary data access request also contains information such as what data and Operation permissions the user can access, and the temporary data access request is valid only for a specified period of time. In this way, the administrator can allow other users to access data without revealing the accesskey pair.

 

2. Computing services

computing Services can be considered as CPU + RAM computing capabilities

Windows Azure provides three Computing Services: Web Site,Cloud Service, and Virtual Machine.

 

 1.Web Site

  Web Site is suitable for lightweight computing.

Web Site features fast and easy deployment of a highly scalable cloud environment. Use your selected language and open-source applications, such as WordExpress,FTP,Git, or TFS, and easily integrate Windows Azure services, such as SQL databases, caches, CDN, and storage.

However, Web Site can only provide basic Windows Azure functions, such as Application and Data. However, more advanced functions, such as Startup Task,Native Code, and Virtual Network, are not supported. Azure Web Site provides four different computing modes:

 (1) Free (Free)

-In the Free mode, your computing resources are shared with other users, not exclusive. In other words, Free Web Site resources are shared with other users.

-You can create up to 10 Free Azure Web Site for an Azure account.

-In the Free mode, the maximum storage capacity of a Azure Web Site is 1GB.

-In Free mode, Azure Web Site does not support scale-out.

-In the Free mode, Azure Web Site is not guaranteed by SLA.

 

 (2) Shared

- in the Free mode, your computing resources are shared with other users, not exclusive. In other words, Shared Web Site resources are Shared with other users.

-An Azure account can create up to 100 Azure Web Site of the Shared type.

-In Shared mode, the maximum storage capacity of a Azure Web Site is 1GB.

-In Shared mode, Azure Web Site supports horizontal scaling and supports up to six Shared instances.

-In Shared mode, Azure Web Site is not guaranteed by SLA.

 

(3) Basic

  -In Basic mode, the customer's computing resources are exclusive.

-In Basic mode, there are three types of exclusive computing resources

(A)Small: 1Core/1.75GB RAM

(B )Medium : 2Core/3.5GB RAM

(C)Large: 4Core/7GB RAM

-An Azure account can create an unlimited number of Azure Web Site of the Basic type.

-In Basic mode, the maximum storage capacity of a Azure Web Site is 10GB.

-In Basic mode, Azure Web Site supports horizontal scaling and supports up to three dedicated instances.

-In Basic mode, Azure Web Site support 99.9% SLA

  

(4) Standard (Standard)

-If the customer's computing resources are exclusive in Standard (Standard) mode

-In Standard mode, there are three types of exclusive computing resources.

(A)Small: 1Core/1.75GB RAM

(B )Medium: 2Core/3.5GB RAM

(C)Large: 4Core/7GB RAM

-An Azure account can create an unlimited number of Standard of the Azure Web Site type

-In Standard mode, the maximum storage capacity of a Azure Web Site is 50GB.

-In Standard mode, Azure Web Site supports horizontal scaling and supports up to 10 dedicated instances.

-In Standard mode, Azure Web Site SLA of 99.9% is supported.

For technical indicators of Azure WebSite, see:

http://azure.microsoft.com/en-us/pricing/details/web-sites/

  

 

2. Cloud Service

Azure PaaS is suitable for newly developed and deployed applications.

  Note that Azure PaaS is a non-persist VM, that is, a non-persistent virtual machine. The non-persistence here means that azure is only responsible for the code developed by customers,

there are two kinds of Azure Cloud Service in a broad sense: PaaS Web Role/Worker Role and IaaS DNS. Here I mainly introduce PaaS Web Role/Worker Role.

  

Azure PaaS allows developers to deploy applications to Azure SDK for platforms based on Microsoft's Azure PaaS. NET,Java,PHP,Node.js,Python,Ruby, etc. The Web Role/Worker Role of PaaS is a very important concept.

  

By using Web Role, developers can use Web Role to deploy HTTP/HTTPS applications, including ASP.. NET, PHP(FastCGI), JSP, or HTTP-based Web applications such as the MQ application. Web Role communicate directly with the outside world.

Worker Role can be simply understood as a Windows Service on the Windows Service, which is an application Role without a user interface and runs silently in the background, developers can use Worker Role to handle a large number of computing logic that does not require a user interface.

Web Role can send a string of messages to Azure Storage Queue in the form of Queue (Worker Role), allowing Worker Role to execute the logic required by users.

 

Why does Microsoft have Worker Role? What are its advantages? Here, let me give you an example. For example, we have an information management system that needs to upload Excel documents for parsing and processing, from the perspective of software design, there are two ways to solve it

(1) create a new upload control in the ASP.NET application and write functions in the upload control: once the Excel file is uploaded, the Excel processing is performed in the. cs file.

However, this has a disadvantage. If the content contained in the Excel file is very large, it takes time to process the content, so the page of ASP.NET in the foreground will be stagnant or unresponsive. Although we can also enhance the user experience by adding progressbar or loading pictures, it is not the best in software design.

(2) the front end still uses the original processing method and uses upload control. A Windows Service is added to the server to query a folder in time series. Once an Excel file is uploaded on the front-end page, the Windows Service process the Excel file. In this way, the front-end page will respond in a timely manner and obtain a better user experience.

However, there is still a defect that the front-end pages and windows service are one-to-one. If the number of attachments uploaded is large, Windows Service may be too late to process them.

(3) with Worker Role, we can make an ASP.. NET page backend has multiple Worker Role for distributed computing. It is a one-to-many relationship that can effectively use cloud computing resources Worker Role handle high-load data access.

 

The advantages of using the third method are as follows:

(1) asynchronous Processing, Web Role only responds to HTTP requests from the client, and responds quickly. The Worker Role processes the messages sent from the Web Role at the backend. The two are loosely coupled.

(2) the relationship between Web Role and Worker Role is many-to-many. For example, I can set Web Role to 10 in the Instance Count configuration. As shown below:

  

In the Worker Role configuration, set the Instance Count to 3

  

the advantage of this is that there are 10 front-end Azure Cloud Service Web Role Instance responding to HTTP requests from clients, while the back-end has 3 Worker Role Instance to handle complex business logic.

 

This structure is like a restaurant with 10 waiters (Web Role) and 3 Cooks (Worker Role).

-The waiter (Web Role) is only responsible for entertaining guests (responding to client requests) and sending the customer's ordering information to the kitchen through Queue messages (Queue.AddMessage())

-The chef (Worker Role) reads the order information (Queue.GetMessage), and then takes charge of cooking (back-end logic). When the meal is ready, the order information is deleted.

-If the front-end pressure is too high (too many client requests), the Web Role can be scaled out.

-If the backend pressure is too high (the backend logic processes too much), the Worker Role can also be scaled out.

-This loosely coupled, many-to-many relationship is ideal for enterprise-level application architectures.

 

    

As you can see, Web Role is an important way to communicate with Worker Role through Azure Storage Queue. Understanding Queue is very important for Azure PaaS architecture design.

Compared with IaaS described below, azure PaaS benefits are summarized as follows:

(1) application-oriented , rather than facing IT infrastructure. As a cloud computing provider, Microsoft allows users to focus more on building excellent software architectures without considering underlying issues, such as networks, operating systems, virtualization, and other IT infrastructures. For example, PaaS(Platform as a Service) allows cloud computing providers to automatically upgrade operating systems and install patches, the installation of patches requires manual configuration and upgrade. The timeliness and reliability are not high. After the PaaS platform is adopted, cloud computing providers and software developers can perform their own duties and focus on problems in their own fields.

(2) elastic . Microsoft Azure Has Worker Role And Web Role. Web Role can respond to frontend events, while Worker Role can respond to requests sent by Web Role. This architecture not only ensures fast front-end response, but also makes the cloud computing architecture more flexible.

 

Azure PaaS platform support

(1) multiple development languages, including. NET, Java, PHP, Node.js, Python, etc.

(2)Azure PaaS virtual machine types, except A0, are exclusive and support the following 8 virtual machine sizes

  

(3)Azure PaaS also supports dynamic horizontal scaling of virtual machines. For example, 10 A7(8core/56GB) instances are processed in parallel.

 

3. Azure Virtual Machine (IaaS)

in terms of cloud computing classification, IaaS (infrastructure as a Service) is oriented to IT infrastructure and services, enabling users to deploy and run their own operating systems, middleware, and runtime. For traditional commercial software, migrating to IaaS platform takes much less time and effort than PaaS. IaaS is more suitable for traditional commercial software.

Azure Virtual Machine Virtual machines are saved in VHD format and stored in Azure Storage Blob. Because Azure Storage supports local redundancy, remote redundancy, and read access regional redundancy, Azure Virtual Machine provides high availability of Storage from the perspective of Storage.

In addition, Microsoft Azure supports uploading Hyper-V hosted virtual machines to Azure cloud for hosting. In this way, for those enterprises that have implemented Microsoft Hyper-V virtualization technology within the enterprise, they can quickly Upload Hyper-V virtual machines within the enterprise directly to Azure, speeding up the migration time.

Azure Virtual Machine supports the following types:

  

for example, Microsoft Azure a single node supports a maximum computing power of A7, that is, 8 cores/56GB. 16 External disks can be attached. The maximum capacity of each disk is 1TB, that is, 16TB of external storage. The maximum IOPS is 16 × 500=8000.

Note: the CPU and memory of the Azure Virtual Machine above are fixed and cannot be changed as you wish.

 

For temporary disks, the disk capacity is related to the virtual machine type. For A7 virtual machines, the temporary disk capacity is 605GB.

For Windows operating system, the drive letter of the Temporary disk is D, and the drive letter name is (Temporary Storage). This temporary disk has the advantages of high IOPS and low latency. However, this disk also has disadvantages: it is not a persistent disk, that is, if the customer saves the files on this disk, there is a risk of losing the files. This disk is actually the disk of the physical Machine where the Azure Virtual Machine Virtual Machine is located. If the Azure Virtual Machine is reset or drifted, the files stored on disk D will be lost. We recommend that you store some non-critical data, such as SQL Server TempDB and Temporary files, in Temporary Storage. http://blogs.msdn.com/ B /wats/archive/2013/12/07/understanding-the-temporary-drive-on-windows-azure-virtual-machines.aspx

  

  

Azure Virtual Machine supports the following Virtual Machine operating systems:

- Windows : Server 2008 R2, Server 2012, Server 2012 R2

-SQL Server : SQL Server 2008 R2, SQL Server 2012 SP1, SQL Server 2014 RTM (Web, Standard, Enterprise)

-Linux :

Ubuntu (12.04 LTS, 12.10, 13.10, 14.04 LTS)

CentOS (6.5)

SUSE (SUSE Linux Enterprise Server 11 SP3)

  

  how to Select Microsoft Azure computing service

if I am an enterprise-level user, which service should I choose to migrate existing enterprise applications to Microsoft Azure cloud computing platform?

The following figure shows the differences between the three services provided by the Microsoft Azure:

  

From the preceding figure, it is not difficult to find that in Microsoft Azure service platform, Web Site features are:

(1) build highly scalable Web sites on Windows Azure.

(2) quickly and easily deploy a highly scalable cloud environment, starting from a small scale.

(3) Use your chosen language and open-source applications, such as WordExpress, FTP, Git or TFS, and easily integrate Microsoft Azure services, such as SQL databases, caches, CDN, and storage.

(4)Web Site is characterized by rapid deployment and can only provide basic Windows Azure functions, such as Application and Data. However, more advanced functions, such as Startup Task,Native Code, and Virtual Network, are not supported.

 

Suitable for Web Site

(1) modern Web applications. Including client-side scripts, server-side scripts, and database applications. And can be scaled out.

(2) continuous development. Deploy data directly from the source code library using Git or Team Foundation.

(3) use open-source applications. You can directly use open-source applications, such as WordPress.

 

Cloud Service features:

(1) resume on the Windows Azure or expand your enterprise application.

(2) use the rich PaaS environment to create highly available and scalable applications and services. Supports advanced multi-layer architecture, automatic deployment, and elastic computing. Through Windows Azure PaaS, powerful SaaS solutions can be provided to customers all over the world.

(3) through Virtual Network, the local LAN Network can be connected to the Windows Azure public cloud Network. In this way, the enterprise network can enjoy the highly elastic computing and interoperability brought by the public cloud, while ensuring network security.

 

Suitable for Cloud Service

(1) multi-layer applications, each layer can expand itself. Using Web Role and Worker Role,Web Role can respond to the front-end display and put complex services on the back-end for processing.

(2) advanced management. If your application requires administrator permissions, remote desktop access, or code running with enhanced permissions, you can use Cloud Service.

(3) private cloud + public cloud. You can use Windows Azure Connect to Connect to the Azure public cloud, or use Azure Virtual Network to Connect the intranet or private Network of an enterprise to the public cloud.

 

Virtual Machine features:

(1) provide IaaS services. Not only Windows but also Linux operating systems are supported.

(2) easily deploy and run Windows Server and Linux virtual machines in a few minutes to migrate the running load without changing the existing code. For Windows OS-based applications, you can deploy them in Hyper-V to make VHD files, and then upload them directly to the Windows Azure for deployment and hosting.

(3) support Virtual Network to directly connect your Lan enterprise applications to public clouds and enjoy the convenience brought by cloud computing.

 

Virtual Machine features:

(1) supports Windows or Linux. You can quickly migrate existing Windows-based or Linux-based applications to Windows Azure.

(2) support more service applications. You can use Windows Azure, MySQL, SQL Server, MongoDB applications in the SharePoint.

(3) supports migrating existing applications to the public cloud. You can save persistent non-relational data to Windows Azure VHD.

 

 

3. SQL Database (SQL Azure)

note that the SQL Database here is different from the traditional SQL Server 2008 R2,SQL Server 2012,SQL Server 2014.

  Microsoft provides Virtual Machine services in Azure Virtual Machine that are consistent with traditional SQL Server. For more information, see Azure Virtual Machine.

 

The SQL Database described here is the SQL Server service of PaaS. In the previous version, it was called SQL Azure, and later it was renamed Windows Azure SQL Database, or SQL Database for short.

In order to distinguish from the traditional SQL Server, the author is used to using the old SQL Azure to call Windows Azure SQL Database, please pay attention to it.

 

In general, if an enterprise needs to create a new database service, it needs to go through the process of purchasing hardware, network wiring, installing operating system, installing drivers, installing database software, etc, the whole process is long and tedious, and IT personnel are required to maintain the database server in the later stage.

If customers subscribe to SQL Azure services, they can use SQL Azure services conveniently and quickly without purchasing any hardware and installation software. To users, SQL Azure is like a Internet Server that has been created on the SQL Server. It is hosted and maintained by Microsoft and deployed in Microsoft's data centers in Shanghai and Beijing. Users can quickly and immediately enjoy SQL Azure services by simply selecting the data center nearest to their physical location.

SQL Azure services:

(1) functions of traditional SQL Server, such as tables, views, functions, stored procedures, and triggers.

(2) Data Synchronization: provides data synchronization and aggregation functions.

(3) management: provides automatic configuration, metering, billing, load balancing, fault tolerance, and security functions for SQL Azure.

(4) Data Access: define different programming methods for accessing SQL Azure. Currently, SQL Azure supports TDS, including ADO.NET,Entity Framework,ADO.NET Data Service,ODBC,JDBC, and LINQ clients.

 

  What is the difference between SQL Azure Database and traditional SQL Server Database? SQL Azure Database provides high availability, scalability, and multi-tenant Database services hosted by Microsoft on the cloud. SQL Azure Database enables independent management, supply, and easier multi-Database deployment. Developers do not have to install or manage any software. For enterprise users, because there is no process of installing hardware and deploying software, the time and cost of obtaining Database are also reduced. For developers, they can use the existing T-SQL development knowledge and familiar relational data models to develop and manage SQL Azure. SQL Azure Database allows us to develop by using existing development tools, such as Visual Studio and SQL Server Management Studio. SQL Azure Database also supports connection methods such as Ado.net and ODBC, and supports Entity Framework.

 

  what are the new features of SQL Azure Database?

  SQL Azure Database automatically performs triple backup, that is, SQL Azure Database automatically copies itself to different physical hosts in the same data center, generating one primary backup and two secondary backups. This improves the reliability, availability, enterprise-level security features, and database security of the SQL Azure. As shown in the following illustration:

  

  

Data Sync ( data Synchronization )

in some special cases, it may be necessary to make the SQL Server data in the Lan consistent with the data in the Windows Azure database on the cloud, SQL Azure Data Sync allows you to synchronize your local SQL Server database Server with SQL Azure databases in the cloud. It supports one-way and two-way data synchronization, therefore, data can be easily shared between SQL Azure database and SQL Server internally deployed database, and between multiple SQL Azure databases in the same data center or different data centers.

 

Use SQL Azure Database what are the benefits?

  (1) reduced total cost of ownership (TCO)

because SQL Azure Database is a relational Database in the cloud, you do not need to install hardware, operating system, Database software, and other processes. Therefore, you do not need IT personnel to manage the Database and do not incur fees such as License; in addition, SQL Azure Database fees are charged based on the number of created databases and the size of the databases. You can delete the Database if you do not need it. This does not incur any fees.

(2) improved availability

because SQL Azure Database supports triple backup, SQL Azure has a 99.9% SLA.

(3) multi-tenant

for independent software developers (ISVs), they can use Web Site when building a set of SQL Azure. By isolating user data and configurations in the same (different) database (data table), multiple users (tenants) can use the same system, in addition, the tenant can only see its own data, but cannot see the data of other tenants (it can also be encrypted, even if other tenants see the data, it cannot be parsed).

 

 What changes have taken place in the development model after using SQL Azure Database?

 

This article is transferred from Lei Zhang blog Park blog, original link: http://www.cnblogs.com/threestone/p/3833397.html,如需转载请自行联系原作者

Selected, One-Stop Store for Enterprise Applications
Support various scenarios to meet companies' needs at different stages of development

Start Building Today with a Free Trial to 50+ Products

Learn and experience the power of Alibaba Cloud.

Sign Up Now