Category Archives: Cloud computing

Facebook faces an insatiable storage appetite

A company that lives on content grapples with storing it all 

TechCrunch is watching Facebook struggle with success.  Their astounding growth in unique visits and page views – up to 35 billion now – is creating a corresponding storage surge that’s sapping their cash.  It’s not just new hardware either.  Their electric bill is estimated at $1 million a month.

The big question is whether they can find an equation where the storage cost per new user is less than the revenue from that user.

Being the biggest isn’t always the best.  Google makes it work.  Facebook has to find their own formula to make their size and growth profitable. And quickly.

Other takeaways:


i365 gets physical with on-site storage

Local appliance addresses the biggest objection to cloud storage: loss of control

i365 has added an on-premises backup and recovery appliance to its EVault online backup and recovery services. 

One of the biggest inhibitors to Cloud Storage for backups has been that businesses don’t want to lose control of their data.  Drunken Data mounted the soapbox on this topic Monday. No matter what assurances a Cloud service makes, it’s hard not to feel safer with data on-site.

The poster child for this reluctance is Amazon S3, which guarantees 99.95% uptime, yet has a history that falls short of this level.

The EVault Express Recovery Appliance stages backups locally, allowing transfers to the Cloud over time.  The incremental costs for the appliance are small compared to conventional 100% on-site backup.  Near-term recovery time is quicker, but maybe more important is the emotional benefit of having recent backups within the company walls (locally or at a remote facility). 

This pragmatic tweak to the Cloud Storage model could open up the business market for SaaS in a big way.  What do you think?

Library of Congress comes to Flickr

Acting on Flickr’s motto: “Share your photos, watch the world”

Helena Zinkham is acting chief of the Prints and Photography Division of the Library of Congress.  Seeing her dedication and passion for history and media as Robert Scoble interviewed her on Scobleizer TV was intriguing  – and comforting somehow, in these crazy days.

What struck me the most: her dedication to her institution as the antithesis of “dead libraries of stuff”.

Turns out the Library has consistently been an early adopter of new media : photographs in the mid 1800’s,  typewriters circa 1900, TV in the 1950’s, the internet in the 1990’s. 

Now it’s participating in Web 2.0 by putting historic photos on Flickr.  Why Flickr?  “Instead of trying to bringing people to the Library, we should go out and participate.”  Helena gets it! 

A core philosophy: represent their objects unaltered, and let other people be their own “Ansel Adams” and optimize the images as they see fit. Given a choice, the Library keeps high-resolution, uncompressed TIF images.  From that they make JPG for easy viewing and GIF for thumbnails.  That’s a lot of bytes per image!

Today there are close to 5,000 LOC images on Flickr – about .04% of the existing collection of 14 million objects. 

I think Flickr’s going to need to add some capacity.

Amazon S3 moves to volume pricing model

Under 50 terabytes? You’re a light user

Power companies charge by the kilowatt-hour; Amazon S3 charges by the GB-month.  Yet another parallel that points to the “utility”ization of storage in the cloud. 

The newest news at Amazon S3 is the addition of a modest volume discount for the really big storers. If you have less than 50 terabytes, you’re a light user and pay full price.  Go figure! 

Om Malik sees the pricing change as a boon for start ups.  He’s right, too, about their amazingly diverse customer base.

Another nugget from Amazon’s S3 post: 70,000 customer objects now stored at the site are touched (stored, retrieved or deleted) per second.  That may sound like a lot, but in any one day, 80% of the 29 billion objects lie undisturbed.  And chances are that the same objects get moved around repeatedly, while a core set (50% ?) rest in peace for long periods.

Amazon S3 may highlight their processing capabilities, but it’s the storage function that pays the bills.

What to do with billowing Cloud Storage

Storage philosophy is as important as storage technology

Storage is growing everywhere, but nowhere more than in the Clouds.  Here’s a link to my guest post on Sitepoint on storage technology and philosophies that can help battle the challenges of billowing Cloud Storage.

Cleversafe’s Storage Internet taking shape

RAIS (redundant array of independent servers) brings new ideas to “can’t lose” data concerns

Cleversafe is applying internet concepts to storage to offer unique value in the SaaS space.  It’s pretty cool – check out Byte and Switch’s take

With five partners now and three more in the wings, they seem to be making it work.  But is it really necessary?  Is the internet broken for storage?

We’ll see how they progress as they take on IBM, HP and EMC.  They see Amazon’s S3 as a potential partner rather than a competitor, with additional redundancy that might shore up S3’s infrastructure availability.