WebProNews

Tag: Amazon RDS

  • Amazon Launches Amazon RDS For PostgreSQL

    Amazon Launches Amazon RDS For PostgreSQL

    Amazon announced RDS (Relational Database Service) for PostgreSQL on Thursday, and that PostgreSQL is now available as a managed service on Amazon Web Services with up to 3TB of storage, 20,000 IOPS and support for high-availability.

    Amazon RDS supports the core PostgreSQL database features, like PostGIS, free text indexing and search extensions.

    Users can scale I/O operations to 30,000 IOPS per database instance. according to Amazon, achieving “consistent, fast performance.”

    Users will be able to deploy production Postgre SQL apps using the multi-availability zone option, and Amazon says RDS will operate a synchronous stand-by replica with an automated fail-over mechanism. It also supports cross-region snapshot copy operations.

    “Since we launched Amazon RDS four years ago, a large number of enterprises and startups have adopted the service because it allows them to run familiar relational databases without any of the operational complexity of on-premise systems, at a substantially lower cost,” said Raju Gulabani, Vice President of Database Services, AWS. “As mobile, web, social and geospatial applications proliferate, we have seen a steady demand for PostgreSQL as a managed service. We’re pleased to bring the agility, manageability and cost benefits of the Amazon RDS platform to PostgreSQL.”

    “Thousands of enterprises rely on ArcGIS to build rich geospatial applications that combine location data with business analytics,” said Marwa Mabrouk, Cloud and Big Data Product Manager, ESRI Inc., which has been testing Amazon RDS for PostgreSQL. “As our customers move their applications to the cloud, many of them choose PostgreSQL. We are excited about Amazon RDS for PostgreSQL because customers can focus on their business and not on the database administration. We think Amazon RDS for PostgreSQL is simple to use, cost-effective and will enable ArcGIS developers to be more productive. We are looking forward to expanding our usage.”

    The product is available around the world in every Amazon Web Services region. More info here.

  • AWS Elastic Beanstalk Now Runs Python Apps

    AWS Elastic Beanstalk Now Runs Python Apps

    Amazon announced that AWS Elastic Beanstalk will now support Python applications, in addition to PHP, Java and .NET.

    Specifically, Elastic Beanstalk will now support Python apps and frameworks that run on Apache HTTP Server and the WSGI interface.

    Amazon also introduced new features, which it says will make it easier to build Python web apps on Elastic Beanstalk.

    “First, you can now easily leverage Amazon Relational Database Service (Amazon RDS) database instances with your Elastic Beanstalk applications,” the company says on the AWS site. “Amazon RDS makes it easy to set up, operate, and scale a relational database in the cloud, making it a great fit for scalable web applications running on Elastic Beanstalk. To learn more about how to setup and use an Amazon RDS database instance with your application, visit “Using Amazon RDS with Python” in the Developer Guide.”

    “Second, you can customize the Python runtime for Elastic Beanstalk using a set of declarative text files within your application,” Amazon adds. “If your application contains a requirements.txt in its top level directory, Elastic Beanstalk automatically installs the dependencies using pip. Elastic Beanstalk is also introducing a new configuration mechanism that allows you to install packages from yum, run setup scripts, and set environment variables. To learn more about customizing your Python environment, visit “Customizing and Configuring a Python Container” in the Developer Guide.”

    To start running Python apps on Elastic Beanstalk, check out Amazon’s developer guide, which includes walkthroughs for Django and Flask apps.

    [Hat tip to CIO.]

  • Amazon RDS Now Supports ASP.NET

    Online retail giant Amazon just announced a new update which allows users to run Microsoft’s database and web programming platform ASP.NET, on its Relational Database Service (Amazon RDS). Amazon reveals that users can now use Microsoft tools they are accustomed to – to deploy, manage and scale SQL Server databases and ASP.NET applications. Amazon also points out that RDS presently runs SQL Server 2008 R2, with an upgrade SQL Server 2012 pending.

    Charlie Bell, Vice President, Amazon Web Services, states, “Amazon RDS now supports three of the most popular relational database engines, giving businesses the flexibility to launch a managed Microsoft SQL Server, Oracle or MySQL database based on the needs of their application. With Amazon RDS, customers no longer have to become experts in database set-up and management and can shift their focus to building and improving the capabilities of their applications. We’re excited to give Windows and ASP.Net developers new options for taking advantage of AWS to quickly deploy and easily manage their databases and applications in the cloud.”

    Some might see this as a tactic to lure users away from Microsoft’s own Azure cloud environment. After all, the two compete – soon after Amazon launched its CloudSearch, Microsoft turned around and adapted Bing search to Azure.

    In related news, a different sector under the Amazonian umbrella just beat out Netflix in a customer satisfaction survey, indicative of possible things to come, as Amazon delves further into the business of streaming content.

  • Cost Effectiveness of Amazon RDS Pay-Per-Usage Software Pricing

    Established software vendors face a difficult balancing act between meeting customer demands for pay-per-usage cloud pricing models while guarding against revenue erosion on traditionally priced offerings. If Amazon’s price for Oracle Database on RDS becomes the norm for price discrimination between traditional and per-per-usage licenses, IT buyers could find themselves paying over a 100 percent premium for the flexibility of pay-per-usage pricing.

    Note, I am only using Oracle as an example here because the pricing of Amazon RDS for Oracle Database is public. This post intends to make no judgments on Amazon or Oracle’s price points whatsoever.

    Pay-per-use software pricing limited to entry level product
    Amazon RDS for Oracle Database offers two price models, “License Included” or “Bring Your Own License (BYOL)”. The License Included metric is fancy terminology for pay-per-usage, and includes the cost of the software, including Oracle Database, underlying hardware resources and Amazon RDS management.

    Three editions of Oracle Database are offered by Amazon, Standard Edition One (SE1), Standard Edition (SE) and Enterprise Edition (EE), listed in order of lowest to highest functionality.

    It’s important to note that pay-per-use pricing is only offered on the lowest function edition, namely, Oracle Database SE1. This should not be a surprise as Oracle, like other established vendors, is still experimenting with pay-per-usage pricing models. Customers can also run Standard Edition One using a BYOL model. This fact, along with Oracle’s list pricing, helps us do some quick and interesting calculations.

    Oracle Database SE1 software price-per-hour ranges between $0.05 to $0.80
    The License Included and BYOL prices both include the cost of the underlying hardware resources, OS and Amazon RDS management. The only difference between the two options is the price of the Oracle Database software license.

    This allows us to calculate the per hour cost of Oracle Database Standard Edition One as follows:

    The Oracle list price for Oracle Database SE1 is $5,800 plus 22 percent, or $1,276 for software update, support and maintenance. Like most enterprise software, customers could expect a discount between 25 to 85 percent. For lower priced software like Oracle Database SE1, let’s assume a 50 percent discount. Although, most customers buying Oracle software are encouraged to enter into Unlimited License Agreements (ULAs) which frequently offer discounts at the higher end of the spectrum.

    All told, Oracle Database SE1 after a 50 percent discount would cost a customer $3,538 (($5,800 + $1,276) x 50%) for 1 year or $4,814 ($5,800 + $1,276 + $1,276 + $1,276) x 50%) for 3 years on a single socket quad core machine like this low end Dell server. Note that Oracle doesn’t use their typical processor core factor pricing methodology for products identified as Standard Edition or Standard Edition One as they are targeted at lower performance servers.

    A single socket quad core machine would offer the performance of somewhere between the Amazon “Double Extra Large DB Instance” and the “Quadruple Extra Large DB Instance”.

    Consider the long term costs of per-per-usage

    Using “Double Extra Large DB Instance” pricing, with our calculated cost an Oracle Database SE1 software license on Amazon of $0.40/hr, we can calculate a 1 year cost of $3,504 and a 3 year cost of $10,512. These figures represent a 1 percent lower and 118 % higher cost of using Amazon’s per-per-usage offering versus licensing Oracle Database SE1 through Oracle for on premises deployment or a BYOL for deployment on Amazon RDS.

    There are obviously multiple caveats to consider, like the ability to get lower or higher discounts from Oracle, or comparing with the “Quadruple Extra Large DB Instance” price point.

    A customer that is unable to get a 50 percent discount from Oracle could save licensing costs by using Amazon’s pay-per-usage offering for Oracle Database SE1. For instance, with only a 25 percent discount from Oracle, the customer could save up to 34 percent on a 1 year basis, but stands to pay an extra 46 percent a 3 year basis.

    Comparing the cost of Oracle Database SE1 using traditional licensing on premises with Amazon’s pricing through RDS, it appears that customers should look hard at Amazon’s per-per-usage offering for up to a 1 year term, but stick with Oracle’s traditional pricing model if the software is going to be used for the typical 3 to 5 year period that companies like to amortize costs over.

    The obvious rebuttal to the above calculations would be that a customer electing for a pay-per-usage model would not necessarily run for 24 hours a day for a full year. While this is true, buyers should understand the long term cost implications before making short term decisions.

    Originally published on rand($thoughts);