AWS Unveils New Machine-Learning Analytics Service

10 Apr 2015 | Author: | No comments yet »

Amazon Cloud Introduces Artificial Intelligence Service.

At the AWS Summit being held in San Francisco today, AWS has just announced that its EC2 container service, a service that developers use to run applications within Linux Containers, is now generally avaialble. Inc.’s cloud division added a new artificial intelligence service that helps users rent technologies to give their applications greater predictive and analytical capabilities.SEATTLE, Apr 09, 2015 (BUSINESS WIRE) — Amazon Web Services, Inc. (AWS), an company AMZN, +0.65% today announced Amazon Machine Learning, a fully managed service that makes it easy for any developer to use historical data to build and deploy predictive models. The company also made a pitch that more and more companies are moving all their computing to AWS, and talked up companies that are doing so, such as publisher Time Inc. These models can be used for a broad array of purposes, including detecting problematic transactions, preventing customer churn, and improving customer support.

This new service will launch into preview in the “near future.” Because it supports the standard NFS protocol, EFS will work with most existing file system tools and applications, so developers can simply mount and manage them with any standard file system tool. It will be interesting to see how Google, the other company in the triad of leaders in the cloud infrastructure market — and arguably one of the most significant machine learning companies in the world — will respond to Microsoft and now Amazon making cloud services anyone can use to train models and make predictions as a managed cloud service. Jassy also announced a new storage service named the Amazon Elastic File System. “A filesystem is the missing building block at the core level of the cloud today,” Jassy said.

Based on the same proven, highly scalable machine learning technology used by developers across Amazon to generate more than 50 billion predictions a week, Amazon Machine Learning’s APIs and wizards guide developers through the process of creating and tuning machine learning models that can be easily deployed and scale to support billions of predictions. According to Amazon, the typical use cases for this service are content repositories, development environments, web server farms, home directories and big data applications — anything, basically, that involves a lot of files. The issue is that the tools to actually analyze data are often expensive and complex – what organizations need is a simple set of tools to do this analytics and machine learning heavy lifting.

He billed it as a “file system that grows and shrinks, automatically.” Among the benefits of the software is that it means customers for cloud computing don’t have to buy storage in advance, they will be billed just for the storage their file systems requirements demand: “With Amazon EFS, you pay only for the storage used by your file system. The move comes as Microsoft and Google also announce services for running containers — a supplement for or alternative to longstanding virtual machines for holding application code on top of physical servers. NAS service that scales across thousands of instances and scales to petabyte scale – storage capacity management is handled automatically by the platform itself.

The service automatically adjusts network throughput and storage IOPS (in out per second), so admins don’t need to worry about that work, Jassy said. Google launched the Kubernetes initiative as an open source Docker container orchestration tool, recently it announced a deep partnership with competing container company CoreOS to tie Kubernetes into CoreOS’s solutions. The announcement builds on Amazon’s partnership with Docker, the startup that made containers a popular technology among developers and increasingly among operations people. Tim Harder, Global Lead, AWS Block Storage Services believes this is a product that will greatly extend the number of workloads that customers will being across to AWS. As I mentioned at the time, this was the beginning of container polarization with Google and CoreOS on one side and Docker and its ecosystem on the other.

Paul Miller, a big data and cloud computing analyst, saw this as less black and white than I did saying that: Google’s Kubernetes and Amazon’s container services are now being more explicit about their ability to support more than just Docker’s approach, and new entrants like CoreOS are offering increasingly capable fresh alternatives. As AWS pointed out in their blog post, the fact that it is increasingly easy to build a data lake is the good news,the bad news is that you need to find data scientists with relevant expertise in machine learning, hope that your infrastructure is able to support their chosen tool set, and hope (again) that the tool set is sufficiently reliable and scalable for production use. Steve Wood, explained customers will be able to use just three steps: “build model; validate & customize; make predictions.” Turning to the database market, Jassy said customers were chomping at the bit for a new database, as shown by attempts to move to technologies such as MySQL. “We get asked all the time is there something you can do that will let us have the same performance characteristics as commercial databases, but as cost-efficient as open databases.” He said the answer was last year’s debut of “Aurora,” a new database cost one tenth the price of offerings from the commercial world. In addition, the traditional process for applying machine learning involves many manual, repetitive, and error-prone tasks such as computing summary statistics, performing data analysis, using machine learning algorithms to train a model based on data, evaluating and fine tuning the model, and then generating predictions using the model. For business owners, the ongoing cost of investing on capital-intensive NAS storage becomes problematic, especially as they’re starting to have a preference to use all-flash storage.

Maybe it’s a case of me seeing battles where none lie – but I can’t help but think that some important players feel very threatened by the rise of Docker. Real time predictions cost $0.10 for every 1,000 predictions plus an hourly reserved capacity charge of $0.001 per hour for each 10 MB of memory provisioned for the model. EFS will take its place among Amazon’s other file storage services, including S3 for object storage, Elastic Block Store for block storage and Glacier for archival storage. And so, Amazon will offer two things “AWS Marketplace for Desktop Applications,” and “AWS WorkSpaces Application Manager.” Jassy said the offerings would change how desktop applications are purchased and managed by companies.

According to an AWS blog post, using the scheduler, teams can handle a number of important requirements: Load Balancing – The Service scheduler allows you to distribute traffic across your containers using Elastic Load Balancing. But with Lambda, each action by a user would spin-up an instance of computing, handle all the scaling and managing, and would only incur costs for each hundreds of a millisecond used. He said customers running Lambda use them to respond to Internet of Things events, databases, and other uses. “It was interesting to us how many mobile developers were using it. To ensure security for this multi-application offering, EFS is built within VPC, AWS’ virtual private cloud that ring fences data for security purposes.

Either way, the announcement from AWS proves one thing, that containers are an increasingly important part of infrastructure and will continue to be so. More than a million customers, including fast-growing startups, large enterprises, and government agencies across 190 countries, rely on AWS services to innovate quickly, lower IT costs and scale applications globally.

Bodell said the company hd moved all its UK publications “100%” to AWS. “Our hosting costs went down from $70,000 a month to $17,000,” said Bodell. The company is guided by four principles: customer obsession rather than competitor focus, passion for invention, commitment to operational excellence, and long-term thinking.

Customer reviews, 1-Click shopping, personalized recommendations, Prime, Fulfillment by Amazon, AWS, Kindle Direct Publishing, Kindle, Fire phone, Fire tablets, and Fire TV are some of the products and services pioneered by Amazon. By the end of 2015, the company will be out of two data centers, will have scaled down a third, and will be “all moved across” by Q1. “The more we do on cloud, it helps us bring in much more agile development … the speed of development has really picked up dramatically, because the staff have so much more power at their fingertips.” Wrapping up the day’s talks, Jassy compared cloud computing to third-party sales of retail goods, which Amazon resisted at first on its own site. It will help them continue their story around migrating legacy data to the cloud – all of those enterprises running NAS systems with spinning disks and looking lustily at all-flash arrays will now have an AWS-style option to go for.

Here you can write a commentary on the recording "AWS Unveils New Machine-Learning Analytics Service".

* Required fields
All the reviews are moderated.
Our partners
Follow us
Contact us
Our contacts

ICQ: 423360519

About this site