Hi all! . In its work, I had to carry out the projects, based both on the same or a different technology, so the idea is to make a comparative description has been brewing quite a long time. After analyzing a large number of similar material on the Internet, and based on personal experience and using one of the Windows Azure MVP, and it was decided to write this post.
A comparison will be organized as a description of the main categories in the field of cloud computing, and describe how one or another category of AWS and implemented in Windows Azure.
development Strategy.
Let's start with the most important aspect I believe - is the development strategy. As we all know today cloud providers are able to provide its services on the 3rd schemes - is IaaS, PaaS and SaaS. Go into a description of these terms I will not because they are already ... Although both platforms make certain spodvizhki towards other schemes, their original ...
Agreement on the level of services provided.
Both platforms provide their customers with SLA `s, ensuring a high level of service ( processing power available for 99.95 % of the time, for other services are guaranteed access to over 99.9% of total time ).
scalability.
How Windows Azure and Amazon Web Services supports automatic scaling. by increasing the amount of RAM ). In addition to the horizontal scaling of computing power Windows Azure can scale databases through technology Federations, the service is available when using SQL Azure. In addition, there are solutions. to simplify the process of building resources. Amazon has a Auto Scaling Service, a Windows Azure - Auto-Scaling Application Block.
additional Features.
In addition to basic services (rental computing power and storage ), both platforms provide a set of additional services. Here, the advantage remains with the Windows Azure, since the list of services is much broader.
AWS. | Windows Azure. |
|
|
computing power.
Both platforms offer a pretty wide range of sizes of virtual machines, ranging from micro dungeons and finishing with an extremely powerful instances. Compare offers from Amazon, and Microsoft can with the following signs.
Amazon Web Services. | Windows Azure. | |
Micro- instance. | Micro instance. | Extra-Small instance. |
The standard offer. | Small, Large, Extra Large. | Small, Medium, Large, Extra Large. |
memory. | From 1.7 GB to 15 GB. | From 1.7 GB to 14 GB. |
processor. | From 1 to 4 cores. | From 1 to 8 cores. |
special Offers. | High-memory, High-CPU, Cluster. | |
Operating Systems. | Linux, Windows Server. | Windows Server. |
data Warehousing.
Both platforms support the popular to date data storage systems, including both standard relational and NoSQL- solutions.
Amazon Web Services. | Windows Azure. | |
Blob / Object Store. | The size of one object in Amazon S3 can not exceed 5 TB. There is no limit on the total amount of data and the number of objects. | The size of one object can not exceed 200 GB ( block blob ) and 1 TB ( page blob ). The size of all objects of the same account can not exceed 100 TB, the number of accounts there is no restriction. * Supported by the geo- replication, which can significantly improve the reliability of data storage. |
tablespace repository. | In Amazon SimpleDB is no limit to the number of records. The tables are organized into ... The number of ... | In the Windows Azure Table Storage There are no restrictions on the number of records. Record may have up to 255 fields and can not exceed 1 MB size. The size of the objects in the same account can not exceed 100 terabytes. No limit on number of accounts. |
relational store.
In this category there is a AWS pretty interesting proposal. The platform offers customers two options relational engine running in the cloud: managed and unmanaged. In the case of unmanaged options available to you to choose from ready-made images (known as Application Machine Images) with the already pre-installed database (supported by IBM DB2, Oracle, MySQL, SQL Server Express, SQL Server Standart, Postgres, Syba). After you select a specific image from a running virtual machine and your DBMS is available. Because the virtual machines themselves are unstable and require external storage you still need to configure Amazon EBS service to periodically reset the backups in the repository. The second approach - run. In this case, a service Amazon Relational Database Service, which is set for you MySQL database or Oracle and platform itself is already assumes the task of updating, replication and storage of backups.
Windows Azure provides a relational SQL Azure Service. The developer does not need to install and configure the server, you can use the same programming model for SQL Server. Communication protocol is the same - TDS. Database size is limited to 150 GB. There is native support shardinga database technology with SQL Azure Federations, which allows you to raise productivity to unsustainable heights. Also supported by a number of support and utility services: SQL Azure Reporting Services, SQL Azure DataSync and others.
high-performance computing.
As a general rule for solving complex scientific, engineering and business problems requires a huge number of concurrent data centers, each of which performs only a part of the work but all together they solve a problem ( an example would be the task of molecular modeling ). As AWS and Windows Azure supports high-performance computing. Amazon provides a special kind of virtual machine, called the Cluster Machine, which lets you combine the instances in the cluster, and perform heavy calculations. Microsoft has, in turn, there is a Windows Server HPC 2008 R2 allows you to integrate your own server with cloud virtual machines in any proportion for distributed computing. In addition, both platforms support and greatly simplify the launch of the popular solutions for today as Hadoop, which allows the analysis of conditions in a cloud of large volumes of data.
Development of solutions.
Both platforms are supported by most modern programming languages , besides considerable investment has been pouring into the development of various SDK, addons and plugins for popular development environments, etc. Dr..
Amazon Web Services. | Windows Azure. | |
Portal Developer. | present. | present. |
Eclipse Tools. | present. | present. |
Visual Studio Tools. | present. | present. |
Java SDK. | present. | present. |
Mobile SDK. | iOS, Android. | iOS, Android, WP7. |
PHP SDK. | present. | present. |
Python SDK. | present. | None. |
Ruby SDK. | present. | present. |
NET SDK. | present. | present. |
Emulators.
The big advantage here is for both Windows Azure for local development provided a very useful and convenient environment ( emulated by the runtime as well as data storage ), as simulating the cloud, which is not the Amazon Web Services.
Condition Monitoring Services.
AWS includes a service Amazon CloudWatch, allowing to collect and aggregirovat detailed metrics from the EC2- instances, load balancers, databases, etc. Dr..
Windows Azure provides a special diagnostic subsystem that allows to monitor, and maintain logs aggregirovat applications, IIS logs, performance counters, statistics, databases, etc..
rate.
Probably one of the most important components of success of a platform. Here, unconditional victory Windows Azure. Independent performance tests were conducted by Compuware using their own tools CloudSleuth, debuted this year. Everyone can view the results for the last month on the site CloudSleuth. https://www. cloudsleuth. net / web / guest / global-provider-view. On this website you can also see the model test cloud. At the time of writing this post the situation was as follows:.
Summarizing all the above can make a summary table of all services on both platforms.
category. | Amazon Web Services. | Windows Azure. |
computing power. | ||
Calculations. | Amazon Elastic Compute (EC2). | Windows Azure Compute. |
high-performance computing. | Amazon EC2 HPS Clusters. | Windows HPC Server. |
data processing. | Amazon Elastic MapReduce (Hadoop). | Hadoop on Windows Azure Compute. |
Auto scaling. | Auto Scaling, Elastic Beanstalk. | Scaling Application Block. |
distributed caching. | Amazon ElasticCache. | AppFabric Caching. |
content. | ||
CDN. | Amazon CloudFront. | Windows Azure CDN. |
database. | ||
NoSQL. | Amazon SimpleDB. | Windows Azure Table Storage. |
A relational database. | Amazon Relational DB Service. | SQL Azure. |
REPORTING. | SQL Azure Reporting. | |
synchronization. | SQL Azure DataSync. | |
federation. | SQL Azure Federations. | |
Expanding. | ||
IaaS deployment. | Amazon Elastic Compute. | VM Role. |
PaaS unfolding. | Elastic Beanstalk. | Web / Worker Roles. |
Mass deployment. | AWS CloudFormation. | |
Marketplace. | ||
Marketplace applications. | Custom App Catalog. | Windows Azure Marketplace. |
Marketplace data. | Windows Azure Marketplace. | |
Monitoring and management. | ||
monitoring. | Amazon CloudWatch. | diagnostic API. |
manual control. | AWS Management Console. | Windows Azure Developer Portal. |
communication. | ||
A simple turn. | Amazon Simple Queuing Service. | Windows Azure Queue Storage. |
The advanced queue. | Windows Azure Service Bus Queues. | |
notification. | Simple Notification Service. | Windows Azure Service Bus. |
mail. | Simple Email Service. | |
network. | ||
traffic management. | Amazon Route 53. | Windows Azure Traffic Manager. |
VPN / Hybrid Cloud. | Virtual Private Cloud, Direct Connect. | Windows Azure Connect. |
load Balancing. | Elastic Load Balancing. | Windows Azure Compute. |
security. | ||
Office of authenticity. | AppFabric Access Control Service. | |
storage. | ||
blob storage. | Amazon S3. | Windows Azure Blob Storage. |
disks. | Amazon Elastic Block Storage. | Windows Azure Drives. |
Import / export of media. | Amazon Import / Export. |
As can be seen in terms of functionality comparable to the platform, so the decision is taken is based on the requirements of the project and clearly understand the difference between IaaS and PaaS principles! .
No comments:
Post a Comment