(download)

iTunes

Stitcher Radio

SHOW NOTES
Host: Adam Lesser
Speakers: David Linthicum and Paul Miller

• Winners and Losers in 2012 and prediction winners/losers in 2013
• The rise of the consumer-oriented cloud
• Managing big data applications in the cloud
• Enterprise and government adoption of the cloud
• Industry-specific clouds
• Crazy prediction for 2013

PREVIOUS GIGAOM PODCAST EPISODES:
A.V. Club’s Kyle Ryan Breaks Down the Year’s Best Podcasts

Instgram’s Twit-storm, Netflix nabs Disney, GMail’s Pretty iPad App

RoadMap re-run, our talk with Instagram’s Kevin Systrom

iTunes 11, When Things Connect, Sun Volt

What Aspiring New Media Stars Should Know About Agents and Managers

Holiday Gadget Gift Guide

War Tweets, Google TV and Nexus 4

Director Jay Duplass on low-fi movies through high-tech

Election Dissection, Ditching DSL and Dumping the iPad

Sandy’s Social, Infrastructure Impact and Forstall

Windows 8 Surfaces, and disruption eruption

iPad Mini, iMac gets skinny

Boxee Cloud DVR, Apple Rumors and Chromebook

Commutist interview: Joy of X author Steven Strogatz

Commutist podcast: Patent trolls, Costco ban and Passbook’s home run

Commutist, meet Nerdist, and interview with Chris Hardwick

T-Metro, Broadband Caps, Remembering Steve Jobs

Apple’s iO-Mess, Dirty Data Centers and Tesla

News from the Mobilize Conference

Paul Tough: How Children Succeed and what you can learn from them

The iPhone 5 Event

Come on, Kindle, Light My 4G Fire

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Survey: How apps can solve photo management
• Sector RoadMap: Social customer service in 2013
• Sector RoadMap: Content personalization in 2013

Karl Freund, the VP of marketing for Calxeda, says the company has shipped about 3,000 nodes and 130 systems although none are deployed in production environments yet. He expects the first production deployments to occur at the end of the second quarter of 2013. But most of the conversation was about how ARM-based systems could be used today in the storage market. Not just for cold storage such as Amazon’s Glacier or Facebook’s photo storage effort, but even for the big storage systems for scale out storage and enterprise class storage appliances. Named customers who are evaluating the systems include Scale.io, Gluster and Inktank, the storage startup backed by Mark Shuttleworth of Ubuntu fame that is commercializing Ceph.

There are more, notes Freund, (pictured) who says that when Calxeda servers make it into production environments, they will likely be deployed first in a storage capacity, as storage customers don’t care if the chips are 64-bit compatible. For now, ARM-based systems are stuck only able to address less memory because ARM only has a 32-bit capable core design. Next year ARM will have a 64-bit capable design and systems will be built around them in 2014 (maybe even late 2013). Calxeda plans its 64-bit capable SoC for 2014.

But Calxeda isn’t waiting and in storage, it’s also not focusing on power consumption — the initial draw for ARM-based servers in the scale out data center. For the storage world, where spinning hard drives tends to suck huge quantities of electricity, adding a low-power has a negligable affect on the consumption of an overall system. However, Calxeda boasts that popping in more of its systems on a chip (SoC) are both cheaper and make for faster information transfer and retrieval.

Its tests show roughly a 4X improvement in IOPs for a rack of Calxeda SoCs versus x86-based systems. Adding Calexeda’s SoCs also cuts complexity because the entire system of processing and networking components are integrated on the SoC, and the terabit-plus fabric between cores also offers more network capacity between cores in a system –the so-called east-west networking traffic.

As the market for scale out computing, storage and networking changes the demands made on IT equipment, Calxeda and others are seeing an opportunity that may have begun in servers and the cloud computing environment, but certainly isn’t stopping there. No wonder Intel is trying to catch up with chips of its own. So far, it’s recently announced new Atom-based chips haven’t made the cut for most customers I’ve spoken with (the lack of integration of the entworking and processing hardware is a problem), but in 2014 it will have a new, integrated SoC as well. Then, the competition will really get interesting.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• ARM: on the road to low-power servers
• 9 Companies that Pushed the Infrastructure Discussion in 2010
• The fourth quarter of 2012 in cleantech

The demand for NoSQL databases is booming with the growth of big data and web applications: something of which Amazon and other tech providers are well aware. Amazon already offers a range of relational database services and SimpleDB, another home-grown, non-relational database, which is easy to use and powerful but has some domain scaling limitations. But NoSQL is built to scale along with unstructured web data and is also suited for the distributed nature of web applications.

Modern “high-volume applications require a simple data model but low latency and high-frequency processing,” Vogels said in an interview. NoSQL is great for social gaming and web applications but is also critical for the big data applications demanded by business today, he said. Such applications require the collection and analysis of not only in-house transactional data, but also information coming in over the transom from many sources — Twitter, Facebook, email, office documents — and in many formats.

DynamoDB uses the sheer scale of AWS to spread a customer’s data across enough hardware to achieve the scale needed and, unlike the other database services, DynamoDB will use solid state drives (SSDs) to assure very fast data access, Vogels said. Here is Vogel’s blog about the news.

DynamoDB has been beta tested for months by select customers, including Reed Elsevier and SmugMug. The company also moved its own Amazon Cloud Drive metadata from an unspecified relational database to DynamoDB before the holidays to handle the anticipated spike in traffic, partially driven by the new Kindle Fire, Vogels said.

Amazon opted to build its own NoSQL because it had so much database experience from its own massive operations, Vogels said.  There are no plans to offer additional, third-party NoSQL options, but Vogels said the company will respond to customer demand.

“In the past, businesses [that used tools like this] had to manage their hardware, their replication, their batching and clustering manually, but with DynamoDB, we’ll take care of all of this for you,” Vogels said. The developer can add or subtract database capacity, transactions per second and other features by turning a dial on the AWS Management Console, he added.

A “perpetual” free tier of the DynamoDB service with storage and a set number of reads and writes per month will be offered, said Adam Selipsky, VP of product management for AWS. Amazon typically offers a try-before-you-buy level of its services, but some of them, like the new free Windows EC2 micro-instances  announced this week, time out after a year.

With the availability of DynamoDB, Amazon shows its willingness to keep rolling out significant new capabilities as demand requires.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Cloud and data first-quarter 2013: analysis and outlook
• Emerging trends in the non-relational database market
• Infrastructure Q1: Cloud and big data woo enterprises