logo

Thông báo

Icon
Error

Chia sẻ
Tùy chọn
Xem
Xem bài viết cuối
Offline admin  
#1 Đã gửi : 13/02/2015 lúc 03:21:11(UTC)
admin

Danh hiệu: Administration

Chức danh:

Nhóm: Administrators
Gia nhập: 23-07-2013(UTC)
Bài viết: 6,110
Man
Viet Nam
Đến từ: Vietnam

Cảm ơn: 10 lần
Được cảm ơn: 2 lần trong 2 bài viết

Danh sách các hạng mục trong Window Azure

The primary location for Windows Azure information is located at http://windowsazure.com. You can find everything there from the development kits for writing software to pricing, licensing and tutorials on all of that.

image: Azure Application Charge Model

Figure 4 Azure Application Charge Model

image: Azure Is a PaaS Offering

Figure 5 Azure Pricing

Azure CapabilityChargeRemarks
Server Usage

Small: $0.12 /service-hour

Medium: $0.24/service-hour

Large: $0.48/service-hour

XLarge: $0.96/service-hour

The roles with active applications determine the charges.

Small : (1.6Ghz), 1.75GB memory (moderate IO capacity)

Medium: (1.6Ghz), 3.5GB memory

Large: (1.6Ghz), 7.0GB memory

XLarge: (1.6Ghz), 14.0GB memory

Azure Blobs and Tables$0.15/GBDaily average measured during each billing cycle. See details on how the charges are computed, as it requires more elaboration.
Transactions$0.01/10K transactionsCreate, Read, Update and Delete into Azure Queues, Blobs and Tables is considered a transaction.
SQL Azure: Web Edition$9.99/month (1GB RDBMS)Metadata of a large application or product catalog of a small e-commerce Web site that sells a few hundred items.
SQL Azure: Business Edition$99.99/month (10GB RDBMS)Useful for medium businesses. Or, by data sharing, it is possible to build applications with large data storage needs.
Azure$0.15/100K message operationsA message operation may be a service bus message, an access control token request or a service management API call.
Ingress GB$0.10/GB ($0.30 in Asia)Only the data transferred in and out of the data center will be billed.
Egress GB$0.15/GB ($0.45 in Asia)Only the data transferred in and out of the data center will be billed.

Azure pricing is straightforward with the one exception of storage used by Blobs and Tables. An account’s Azure Storage usage is measured each day during a billing cycle and a daily average is computed. The charge will be computed by multiplying this daily average by $0.15/GB. For example if you store 20GB on day one, add 10GB on day two, add 5GB on day three, and delete 5GB on day four, with no activity during the rest of the billing cycle, the price will be computed as shown below:

((20 +10 + 5 – 5)/30) * 0.15 = $0.15

This assumes a 30-day billing cycle. Daily sampling of storage will make sure that applications with highly transient storage needs will still pay for their storage usage, unlike a system that measures only at the end of the billing cycle.

As mentioned earlier, architecture is an important factor in the monthly operating cost of an application. For instance, if an application generates lots of data and only the latest data—say the last two weeks—is needed for the functionality of the application, the architecture can be tweaked to delete the unneeded data or to periodically transfer it to on-premises systems. You may be better off by paying a onetime bandwidth cost than incurring perpetual storage costs. The same can be true with the reference data that is no longer part of the active data set. This approach may work well for companies that have already invested in data archival capacity.

The Application Scenario

I will look at various aspects of Azure in the context of an industrial-strength e-commerce scenario: the Shopping List application. I will focus on creating a grocery list and saving it for later use while shopping at the store. The Web UI composes the shopping list and uses Web services to save it to Azure Storage. For scalability, the Web tier writes shopping lists to Azure Queue; periodically, a batch process polls the lists from queues and saves them to Azure Tables. I will use Azure-based authentication and role-based security to demonstrate real-world solution aspects.

image: Shopping List Application on Azure

Figure 6 Shopping List Application on Azure

In the context of the cost-oriented architecture discussed previously, various decisions will impact monthly operational expenses. Here are a few aspects of a system that must be considered before proceeding with the architecture:

  1. Growth rate of reference data
  2. Growth rate of transactional data
  3. Capture rate of behavioral profiling data
  4. Growth rate of business event data
  5. Capture rate of system event data
  6. Media content related to products
  7. Using queues vs. direct interaction with persistent storage

My Shopping List scenario didn’t include much media content, so that wasn’t a big factor in the cost equation, but it may be very important to consider for content sites that deliver videos, imagery and audio streams. Figure 7 shows for a typical application the monthly operational cost on Azure. The spreadsheet doesn’t include the personnel costs for development, operational support and end user support.

image: Azure Operating Expenses Calculator for an E-Commerce Application

Figure 7 Azure Operating Expenses Calculator for an E-Commerce Application

A cloud computing environment will reduce the number of operational support staff, so this should be factored in when comparing the ROI between on-premises and the cloud. Also, it’s important to include power and depreciated capital expense per application in the equation for ROI. Current on-premises application cost models often don’t include these expenses, as it’s very difficult to break down the power consumed on a per-application basis. The same is true for cooling and floor space. ROI calculators can use educated guesses in the absence of objective cost breakdowns.

The simple cost calculator shown in Figure 7 estimates the operating expense of applications hosted on Azure. This Microsoft Excel-based tool allows various input parameters of a typical e-commerce application, and it computes the monthly operational cost using the Azure pricing table shown in Figure 5. Please keep in mind that the default parameters used in the tool are fictitious; you need to take your own system into consideration before making decisions based on the tool output. The cost calculator is driven by the number of visitors per month and assumes a certain number of page views and transactional and event data creation. The Azure team created a more comprehensive tool that calculates the monthly cost of an application and also compares the TCO of on-premises applications with that of Azure.

As shown in Figure 7, our fictitious application generates 9000GB of data in a given month, which costs about $1,350 per month if we were to store this inside Azure Tables. Please keep in mind that Figure 7 only shows point-in-time storage, and event-data charges can accumulate as the application continues to operate. Such costs can be optimized by tuning the amount of event data captured as an application matures operationally. The cost calculator is driven by the number of visitors per month and uses a hypothetical number of 10 Web roles and 3 worker roles. The total monthly bill is $3,571.

Alternatively, the application can be architected to channel the event data by paying onetime bandwidth costs ($0.10/GB transferred out) to an already-depreciated on-premises storage system, if it exists. Similar strategies can be applied to transactional and behavioral profiling data to avoid cumulative storage charges.

Compute charges aren’t cumulative in nature and thus have less impact on the overall operational expenditure of the application. However, there is opportunity to tune the number of active Web and batch role instances based on the observed scalability profile of the application, to get marginal relief on the operating expenses. Between compute and storage charges, compute usage can be controlled at any given time, whereas storage cost depends on architectural decisions that can’t be undone easily once the application is built. So my suggestion is to get your persistence architecture right the first time.

In addition to the cost model of the application, large enterprises will pay close attention to application security, which I will explore now.

Compute Security

Enterprises are finicky about application and data security in the cloud. While security of the datacenter, infrastructure and the operating system are taken care of by Microsoft, application security is still the responsibility of the application owners. I will look at application security from the perspective of my Shopping List Web application. Securing Azure application is similar to its on-premises counterpart. Azure provides various system components to help developers integrate security into applications. These system components allow basic, self-contained authentication and authorization to federated scenarios suitable for large enterprises.  

Basic Identity

A basic identity model, as the name suggests, implements a self-contained identity architecture to meet the needs of one application or a co-located group of applications that share the same set of users and are tightly coupled to the same identity system at the implementation level. The Azure samples contain a set of ASP.NET providers (membership, role, profile and session) that can be used for implementing a basic identity solution. Azure ASP.NET providers are implemented on Azure Storage, which includes Azure Tables and Blobs. These providers implement the ASP.NET provider contracts and leverage StorageClient APIs that are part of the Azure SDK. The schematic of the providers is shown in Figure 8.

image: Azure ASP.NET Providers

Figure 8 Azure ASP.NET Providers

In order for the applications to use Azure ASP.NET providers, the Web.config file needs to be modified to remove the default providers and include new ones. The configuration changes shown in Figure 9 are similar to the changes that must be made for custom ASP.NET providers in on-premises situations.

Figure 9 Web.config Changes for Azure ASP.NET Providers

<system.web>

... ... ... ...

<authentication mode="Forms" />

<!-- Membership Provider Configuration -->

<membership defaultProvider="TableStorageMembershipProvider"

userIsOnlineTimeWindow="20">

<providers>

<clear/>

<add name="TableStorageMembershipProvider"

type="Microsoft...AspProviders.TableStorageMembershipProvider"

description="Membership provider using Azure storage"

applicationName="ShoppingList"

... ... ... ... ...

minRequiredNonalphanumericCharacters="0"

requiresUniqueEmail="true"

passwordFormat="Hashed"/>

</providers>

</membership>

<sessionState mode="Custom"

customProvider="TableStorageSessionStateProvider">

<providers>

<clear />

<add name="TableStorageSessionStateProvider"

type="Microsoft...AspProviders.TableStorageSessionStateProvider"

applicationName="ShoppingList"/>

</providers>

</sessionState>

<roleManager enabled="true"

defaultProvider="TableStorageRoleProvider"

cacheRolesInCookie="true"

cookieName=".ASPXROLES"

cookieTimeout="30"

... ... ... ... ...

cookieProtection="All">

<providers>

<clear/>

<add name="TableStorageRoleProvider"

type="Microsoft....AspProviders.TableStorageRoleProvider"

description="Role provider using table storage"

applicationName="ShoppingList" />

</providers>

</roleManager>

... ... ... ...

</system.web>

Once the ASP.NET providers are configured, authentication, authorization and user profiles can be implemented similarly to traditional ASP.NET applications. Note that the configuration in Figure 9 contains Azure storage-based session provider, which allows the storage of session states on a durable medium. Since Azure load balancers don’t support sticky sessions, storing session data on Azure storage offers a better user experience through session-based personalization. The basic identity model is suitable for applications that have user identity lifecycles (user account creation, usage and closure) that begin and end in the same application. Basic identity implementation may be elected for a variety of reasons, including the following:

  • An application wants to retain the complete ownership of the user identity records
  • A lack of infrastructure for implementing federated identity store and supporting services
  • Applications with a short lifespan (for example, marketing contests and promotions) that require user registration

Azure ASP.NET providers can also be used to authenticate users from AJAX as well as Silverlight applications. The AJAX callable AuthenticationService, ProfileService and RoleService classes, located inside System.Web.Extensions.dll, can be published as .svc endpoints through the Azure Web role. Keep in mind that these services require ASP.NET compatibility for accessing HTTP context-specific data. The article titled “Build Line-Of-Business Enterprise Apps with Silverlight, Part 2”(msdn.microsoft.com/magazine/dd434653), gives detailed  information on setting up the above services to be called from Silverlight or AJAX.

Federated Identity Model

Federated identity is necessary for applications that include supply chain, value chain, collaboration and social networking, as well as applications that integrate popular identity stores on the Internet. The Azure ASP.NET stack can be combined with Windows Identity Foundation (WIF) to integrate with one or more security token service providers. WIF works in conjunction with the pre-established trust relationships enabled by WS-Trust and WS-Federation. Figure 10 shows a conceptual view of the Shopping List application working with two token providers—one on-premises and the other a fulfillment partner.

image: Multiple Token Services Can Be Registered with Azure Applications

Figure 10 Multiple Token Services Can Be Registered with Azure Applications

The trust describes the Secure Token Service (STS) endpoints and the necessary X509 certificates for signing token requests and responses. Figure 11 shows the trust schematic, the XML representation of which will be included in the Shopping List application configuration at the time of deployment. Users get authenticated in their respective systems and the resulting Security Assertion Markup Language (SAML) token gets forwarded to the requesting application.

image: Federated Trust Descriptor

Figure 11 Federated Trust Descriptor

As shown in Figure 10, when a user accesses secure Web content on the Azure-hosted Shopping List application, WIF forwards the request to a Shopping List STS URL present in the trust configuration. The Shopping List STS gathers credentials, authenticates users against Active Directory, constructs a SAML token with the help of Active Directory Federation Services (ADFS, formerly “Geneva Server”) and forwards it to the Shopping List application via the Web browser. WIF running inside the Shopping List site on Azure will extract SAML claims and perform authorization checks.

When multiple STSes are involved, a Web site will have to implement token translation logic for converting diverse tokens into a canonical format. To minimize the impact of introducing a new STS into the system, the token translation logic can be externalized or encapsulated into a component that can be modified without impacting the applications that consume them. Figure 12 shows the token translation schematic that works in conjunction with WIF.

image: Federated Identity System with Multiple Token Providers

Figure 12 Federated Identity System with Multiple Token Providers

Scenarios such as the following will be enabled by the federated identity model:

  • Storage of identity records on-premises for regulatory compliance
  • Leveraging the existing on-premises application security infrastructure
  • Integrating with partners in the value chain and supply chains
  • Single sign-on between the on-premises and the Azure application

Often, large enterprises have already implemented authentication services and directory servers that need to be leveraged for securing applications. Azure allows leveraging of the cloud for expedited application deployment while at the same time taking advantage of the existing infrastructure for security. Also, Azure by design allows the use of federated identity that enables various integration scenarios across business partners and value chains.

Azure Storage

Applications and services deployed on Azure may use Azure Storage for persistence of unstructured and semi-structured content. Azure Storage comprises three fundamental capabilities necessary for building industrial-strength applications and services: Tables, Blobs and Queues. Azure Storage is a massively scalable and highly reliable persistence mechanism that is also accessible to the applications hosted on-premises through industry-standard Web services interface like REST. For on-the-wire privacy, Azure Storage supports SSL (HTTPS)-based access in addition to the standard HTTP protocol. Scalability and other systemic qualities are achieved through a large storage farm comprising commodity server hardware and disk arrays, which is managed by Azure Storage software. The storage access is load-balanced automatically across a set of nodes, for scalability and availability. Each node is responsible for a finite amount of physical storage. Access to storage outside a node’s scope is accomplished through a peer-to-peer interface. The reliability is achieved through the redundancy of the stored entities (such as ShoppingList) on multiple nodes. The storage software makes multiple replicas (three at the time of this writing) of the data automatically once a write occurs. Storage supports atomic transactional writes, and the transaction will complete only after all the replicas are written to the drives. Figure 13 shows a collection of commodity storage nodes forming the Azure Storage Service.

image: Storage Service

Figure 13 Storage Service

While being used, any storage drive anywhere may fail, the possibility of which is shown by the red “X” on node numbers 4 and 11. Once the storage service identifies a failed drive, it replicates the data from a functioning drive to a new node. The storage service is always compliant with the replica policies at any given point in time. As mentioned earlier, the request traffic from applications will be load-balanced across multiple nodes.

This kind of architecture will help the massive scales required by public cloud PaaS offerings such as Azure. As shown in Figure 13, let us assume that nodes 4, 11 and 14 own the initial three replicas of a piece of data. In the event of the failure of nodes 4 and 11, node 14 will continue servicing the requests directly, as well as quickly re-replicating the data to at least two additional nodes (node 2 and 8) to keep the data at a healthy number of replicas. 

Storage Security

Azure Storage relies on Hash-based Message Authentication Code (HMAC) for authenticating the REST Web requests. The shared, secret key associated with the Azure Storage project is combined with the HTTP request in computing a 256-byte hash that gets embedded as an “authorization” header into the Web request. The same process is repeated on the server to verify the authenticity of the request. Azure Table, Queue and Blobs all follow the same authentication process, while the payload and the target URLs are different for each of the storage types. The following are the URLs for accessing the above three storage capabilities under the project, say “hkshoppinglist”:

  • http(s)://hkshoppinglist.blob.core.windows.net/
  • http(s)://hkshoppinglist.queue.core.windows.net/
  • http(s)://hkshoppinglist.table.core.windows.net/

The code sample in Figure 14 shows the creation of multiple Azure Tables as a part of the storage preparation for application deployment.

Figure 14 Pseudo-Code that Shows the Authenticated Creation of Azure Tables and Data

[DataServiceKey("TableName")]

Public class StorageTable

{

Private string _tableName;

Public string TableName

{

get { return this._tableName; }

set { this._tableName = value; }

}

}

Public class Customer: TableServiceEntity

{

Public string Name { get; set; }

Public string CustomerID { get; set; }

public Customer()

{

PartitionKey = "enterprise";

RowKey = string.Format("{0:10}_{1}", DateTime.MaxValue.Ticks –

DateTime.Now.Ticks, Guid.NewGuid());

}

}

CloudStorageAccount _storageAccount = CloudStorageAccount.FromConfigurationSetting("DataConnectionString");

Public void CreateMultipleCustomers(List<Customer> customers)

{

TableServiceContext tsc = new

TableServiceContext(_storageAccount.TableEndpoint.AbsoluteUri,

_storageAccount.Credentials);

foreach (Customer cust in customers)

{

tsc.AddObject("customers", cust);

}

try

{

DataServiceResponse resp = tsc.SaveChanges(SaveChangesOptions.Batch);

foreach (ChangeOperationResponse cor in resp)

{

if (cor.Error != null)

{

//cor.Headers["Location"] can be parsed to find out the failed

//requests which can be retried after correcting the error condition

}

}

}

catch (Exception ex){ //do something with the exception }

}

protectedvoid linkCreateTables_Click(object sender, EventArgs e)

{

labelStatus.Text = string.Empty;

try

{

CreateTable("customers");

CreateTable("products");

}

catch (DataServiceRequestException ex)

{

labelStatus.ForeColor = System.Drawing.Color.Red;

labelStatus.Text = "Error: Table creation error : " + ex.Message;

}

}

//Use ADO.NET services directly to create an Azure Table

Public void CreateTableUsingContext(AzureStorageTable storageTable)

{

TableServiceContext tsc = new

TableServiceContext(_storageAccount.TableEndpoint.AbsoluteUri,

_storageAccount.Credentials); tsc.AddObject("Tables", storageTable);

try

{

DataServiceResponse resp = tsc.SaveChanges(SaveChangesOptions.None);

//handle errors

}

catch (Exception ex){//do something here}

}

//much simpler way of creating an Azure Table

publicvoid CreateTable(string tableName)

{

CloudTableClient ctc = _storageAccount.CreateCloudTableClient();

try

{

ctc.CreateTable(tableName);

}

catch(Exception e) { //handle exception }

}

Using Azure Tables as an example, I will show some simple ways of preparing Azure Storage for transactional population of data. The code samples show the creation of “customers” and “products” tables using TableServiceContext as well as CloudTableClient to illustrate the flexibility of the REST-based interaction. In fact, you can also craft a raw payload, attach HMAC to the Web request and do an HTTP POST to the table URL, but it requires lot of code and should only be done as an academic exercise. The recommended approach is to use StorageClient, which is part of the Azure SDK.

The CreateTableUsingContext function uses the AzureStorageTable class to generate the table creation payload with the help of ADO.NET Data Services. TableServiceContext automatically generates HMAC and attaches to the request using the key contained in the CloudStorageAccount.Credentials property.

Azure Table storage allows batch transactions, as shown in the function CreateMultipleCustomers in Figure 14. The batch should not exceed 100 operations in a given change set, and a single batch should not exceed 4MB in size. For more details, please refer to the Azure Storage documentation. Batch transactions are only allowed with the entities belonging to the same partition.

Credentials necessary for the generation of HMAC are specified in the service configuration of the respective Azure role. The following is the format of the connection string for local storage and the cloud:

Local storage:

<Setting name="DataConnectionString"

value="UseDevelopmentStorage=true"/>

Cloud storage:

<Setting name="DataConnectionString"

value="DefaultEndpointsProtocol=

http;AccountName=

<your account>;AccountKey=<your account key"/>

There is no notion of role-based security in Azure Storage; so an authenticated request will have complete access to the storage in the context of the storage project. An exception to this is the blob container, which can be public (anonymous) or private. Authorization is the responsibility of the application that consumes the storage services.

Wrapping Up

In this article I only scratched the surface of Azure. I am sure there will be plenty of coverage in the future about Microsoft SQL Azure, Azure, various server roles and other security scenarios not covered here. Azure is a cloud computing platform that is architected to enable on-demand utility computing for developing and hosting applications and services.

Large pools of commodity hardware are made highly reliable by software through a high degree of automation. The economic advantages of massive scales are passed back to consumers through a low subscription fee. Subscribers will be charged based on the usage of bandwidth, storage and compute cycles over a monthly billing cycle. Azure comes with the platform components necessary for building enterprise-class applications and services with no upfront commitment of capital or long-term contracts.

What Is Windows Azure IaaS?

Some of you may have started down the path of working with cloud computing and have run into the phrase “Infrastructure as a service” or IaaS. You might be wondering what IaaS can provide for you and your business. You may even get a little confused about IaaS and all the similar terminology that is out there: PaaS, SaaS, STaaS, SECaaS, etc. In this article we are going to take a brief look at IaaS, then we will take a longer look into Microsoft’s Azure IaaS and how you can get started with Windows Azure IaaS.

Cloud Services

To understand cloud, think quickly on your on premise datacenters. In your datacenters you manage everything from the networking and servers to the applications and everything in between. When you begin to leverage clouds services you offload some or all of the management of your datacenter services. To understand this, let’s take a quick look at two other main cloud services: Platform as a Service (PaaS) and Software as a Service (SaaS).

Cloud services explained

PaaS, SaaS, and Iaas

PaaS provides a platform for your applications, and you have really no control over the platform. With PaaS you only had to manage the applications and data, and the cloud service managed the rest of infrastructure. This is by design: PaaS was the original offering of Windows Azure and provided a highly scalable Windows platform for your applications. You still architected, created, compiled, and tested your applications to run in Windows Azure. You then uploaded the application and configured or provisioned your storage.

With SaaS, you are not responsible for supporting the application or any of the components. You are basically paying a provider for a service, usually on a subscription basis. With SaaS, your provider manages the data and platform — really the entire backend. SaaS is probably the most prevalent of the cloud services. Some of the SaaS services include Office 365, CRM, Salesforce.com, and others. Additionally there are plenty of free SaaS services like XBOX Live.

That brings us to IaaS. Azure IaaS gives you a combination of control and the ability to leverage the cloud infrastructures. With IaaS, you manage the application, data, runtime, middleware, and even the OS. The main part with the virtualization and hardware are managed and maintained by your cloud provider. You will still perform many administrative duties in IaaS, and you will still have to install applications, manage patches, manage virtual machines and networks used by your virtual machines, etc. You configure the environment how you want it, and at the end of the day it is yours! With Windows Azure IaaS Microsoft takes the responsibility for uptime of that machine. For example, if a machine crashes, Microsoft will spin up a new instance of your server for you.

It’s easy to get IaaS and PaaS confused. Just remember: If it requires a developer, it’s not IaaS.

Microsoft is the only vendor to provide offerings in all three main types of cloud computing services: IaaS, SaaS, PaaS. Microsoft provides a full spectrum of public cloud solutions, and has a strategy of integrating with traditional on-premises datacenters for Hybrid scenarios. Microsoft does not limit you to go with a 100-percent private or public cloud. This may give you the flexibility to choose the best cloud model based on your unique organizational and application needs.

Danh sách các hạng mục trong Window Azure

1. COMPUTE (Virtue Machine, Web Sites, Mobile Services, Cloud Services)

2. DATA SERVICES (Storage(Blobs, Tables, Queues) , SQL Database, SQL Data Sync, HDInsight, Hyper-V Recovery Management, Backup, Cache)

3. APP SERVICES (Notification Hubs, Service Bus, Media Services, BizTalk

Services, Active Directory,  Scheduler, Content Delivery Network,

Multi-factor Authentication, Automation,  Visual Studio Online )

Windows Azure IaaS

When Windows Azure was first launched, the focus was squarely on PaaS and providing a platform for applications. As Windows Azure evolved Microsoft added many new functionalities and support for other application languages. In particular Microsoft added a function that some of you may be familiar with called the VM Role. The VM Role allowed you to deploy custom Windows Server 2008 R2 images to Windows Azure. This provided a way for you install squirrely apps, maybe error prone apps, or ones that do not install easily or quickly. This is not IaaS — the VM role was still PaaS. While you still built your own image, it had to conform to the PaaS model. More importantly, the VM role was stateless and had non-persistent local storage; you would need to leverage Windows Azure storage service for persistent storage. You can learn more about the VM role here on the Overview of the Windows Azure VM Role.

And then in steps the next evolution of Windows Azure: Virtual Machines. Virtual Machines introduces functionality that allows full control and management of virtual machines and networking. Virtual Machines provides persistent and easily extended storage. This allows you to build your virtual machine directly in Windows Azure or upload and virtual machine you created locally. This provides many options not only for your applications that require persistent storage but also many other infrastructure opportunities. This provides the answer to Microsoft customers for this common for Windows Azure to provide this functionality. This is truly IaaS.

Windows Azure virtual machines

Creating virtual machines in Windows Azure

Virtual Machines provides a true solution for IT professionals, and your responsibilities are almost the same for your on-premise datacenters compared to the machines hosted in Windows Azure. You will still have the considerations to configure your virtual machine, its storage network access, and applications you want to install. Windows Azure Virtual Machines provides you the ability to run many of today’s key server applications such as AD, DNS, SharePoint. The functionality is there for you leverage. You will still have to perform your daily administration of these services, but the main difference is that you do not have to worry about the hardware or high availability. When you create, or upload, a Virtual Machine in Windows Azure, the Windows Azure service creates a replica copy for you automatically on the backend.

Creating a VM in Windows Azure is pretty straight forward, as you choose among these predefined VM sizes and capacity categories. The different sizes — from extra small to extra large — from which you can choose will determine many aspects of your virtual machines, whether it be the number of CPU Cores you have dedicated for your use (or shared in the case of extra small) or the amount of memory, bandwidth, and number of Data Disks preconfigured for you. You will even have the ability to include data disks that are separated from the OS disk. Each of the persistent storage disks you create can be up to 1 TB in size. Since this is still in preview, the numbers and sizes of the virtual machines may change when the service is released.

Hybrid Infrastructure and Iaas

One aspect that makes IaaS great is the ability to leverage IaaS as part of overall architecture and offer the ability to extend your on-premise data centers. I know what you’re thinking: Wait, I thought I had to put it all in the cloud or nothing. That is simply not the case. This is a common use of IaaS, which allows you to straddle the fence between on-premise datacenters and cloud services. You can have part of your application on-premise and the rest in the cloud. This is a common example of a Hybrid Infrastructure or Hybrid Cloud. IaaS allows you to have the best of both worlds for your infrastructure requirements and business demands. It provides native capabilities for you to quickly network these environments together. In IaaS you can leverage private networks for just your VMs or extend your own datacenter into the cloud so that the machines have full participation in your LAN and/or your extended datacenter resources.

Ai đang xem chủ đề này?
OceanSpiders 2.0
Chủ đề tương tự
Danh sách địa điểm bán khẩu trang trên địa bàn TP.HCM (Tin trong nước)
Bởi Ellry 28-03-2020 lúc 05:42:55(UTC)
Danh sách các tổ chức từ thiện nổi tiếng tại TP. Hồ Chí Minh (Văn hóa - xã hội)
Bởi admin 24-08-2019 lúc 12:43:51(UTC)
Danh sách 21 trường có yếu tố nước ngoài tại TP.HCM (Tin trong nước)
Bởi Ellry 09-08-2019 lúc 09:44:01(UTC)
Những việc cần làm để chốt danh sách khách mời dự tiệc cưới (Kế hoạch chuẩn bị cho đám cưới)
Bởi Ellry 16-10-2018 lúc 05:48:00(UTC)
Lưu ý khi làm danh sách khách mời đám cưới (Kế hoạch chuẩn bị cho đám cưới)
Bởi Ellry 08-06-2018 lúc 11:34:37(UTC)
Danh sách miễn Visa, Visa xin cửa khẩu và thị thực điện tử áp dụng cho hộ chiếu Việt Nam (Du lịch nước ngoài)
Bởi Ellry 19-04-2017 lúc 10:53:53(UTC)
Hướng dẫn kiểm tra xử lý khi IP bị liệt vào danh sách Blacklist (Mail Server)
Bởi admin 03-11-2016 lúc 05:51:25(UTC)
DANH SÁCH TOP NHÀ HÀNG, QUÁN ĂN TRƯA NỔI TIẾNG ĐÂT SÀI GÒN (Món ăn Châu á)
Bởi anuongthaga 17-10-2016 lúc 10:56:04(UTC)
Danh sách nhà hàng buffet dimsum, đồ thái, đồ âu hàng đầu tp.hcm (Món ăn Châu á)
Bởi anuongthaga 12-10-2016 lúc 09:24:11(UTC)
Tổng hợp danh sách những nhà hàng uy tín tại Hà Nội (Thế giới ẩm thực)
Bởi Vui Nguyễn 08-10-2016 lúc 11:33:05(UTC)
ĐIỂM QUA DANH SÁCH HƠN 80 NHÀ HÀNG BUFFET NGON VÀ KHÁ CHẤT LƯỢNG TẠI SÀI GÒN (Món ăn Châu á)
Bởi anuongthaga 27-09-2016 lúc 02:14:02(UTC)
DANH SÁCH 60 NHÀ HÀNG HÀNG ĐẦU HÀ THÀNH CHO DÂN SÀNH (Món ăn Châu á)
Bởi anuongthaga 16-09-2016 lúc 04:41:34(UTC)
Lưu lại danh sách các nhà hàng ngon để đỡ mất công tìm nhé (Thế giới ẩm thực)
Bởi Lan Quỳnh 14-09-2016 lúc 02:47:13(UTC)
Danh sách công cụ kiểm tra backlink cần thiết (SEO)
Bởi admin 31-08-2016 lúc 06:32:39(UTC)
Danh sách các loại rau củ quả màu tím cực tốt cho sức khỏe (Thực phẩm chức năng - dinh dưỡng)
Bởi Ellry 30-08-2016 lúc 09:59:00(UTC)
Di chuyển  
Bạn không thể tạo chủ đề mới trong diễn đàn này.
Bạn không thể trả lời chủ đề trong diễn đàn này.
Bạn không thể xóa bài của bạn trong diễn đàn này.
Bạn không thể sửa bài của bạn trong diễn đàn này.
Bạn không thể tạo bình chọn trong diễn đàn này.
Bạn không thể bỏ phiếu bình chọn trong diễn đàn này.

| Cung cấp bởi YAF.NET 2.2.4.14 | YAF.NET © 2003-2020, Yet Another Forum.NET
Thời gian xử lý trang này hết 1.538 giây.