Monkey Analytics Blog

Now that we have actual users (and therefore actual testimonials and actual user stories) we’ve updated our website accordingly, and retired our original “testimonials” (which were obviously self-generated).  For posterity, we’re posting the old “testimonials” here:

Best web computation software available!

Of course, I’ve never used any such software, and don’t know what to do with it. But I do know how excited this team is to have built this tool, and I may be a bit biased too.

~Mother of company founder

Ook, ook, uk, ooga

Before Monkey Analytics, I was just another monkey in the cage at the zoo, free (well, not really) and happy. But now they make me perform math computations all day long, and have restricted my diet of bananas to force me to work harder. As a protest I return all my answers as 42.

~Disgruntled monkey

This has been a long time coming

I’ve always wondered why fantastic tools such as Matlab didn’t focus more on the usability aspects of their products. That left a hole wide open for a company such as Monkey Analytics to fill. Of course, Matlab could never replicate their shrink wrap license revenues as a service company, so perhaps it is best that Monkey Analytics is around to provide this service affordably.

~Random engineer

You asked and we listened.  Preliminary support for sessions running The R Project code just went live.  And sorry, no we do not yet support CRAN installs - see our earlier post on the matter – but we will install whatever packages you request!

Monkey Analytics has achieved the trifecta – running the Matlab M language, Python, and The R Project all in the cloud.  Pretty darn cool toolbox if we do say so ourselves!

More to come later , but we wanted to get this early functionality out for you to play with and critique.  Enjoy!

At another user’s request, we have installed the Octave-Forge packages. We know many of you would like the ability to install packages yourself, but we don’t have a secure system to allow this yet (sorry – more details in our FAQ here).

In the interim, we will install packages for all users to use as part of our base computing package.  If your favorite package isn’t there, let us know and we’ll work to add it.

Keep hitting us up with the feature requests – and we’ll keep coming up with solutions so that you can compute away!

We launched a new feature last night – you can now upload your data files to the system. Have a favorite .wav file you’re crunching on? An image to process? Signed in users will notice a new link in the data section titled “Upload data”. Upload away!

We also have a ‘process’ for larger data sets – contact us and we’ll work to ftp / scp larger data sets directly into your Monkey Analytics user data area.  (Longer term plans should automate this, but at the current startup-stage, this solution works).

A few bug fixes too – not worth mentioning details, but we know some of you noticed glitches – we did too – and have worked to improve the product.

We love finding more bloggers talking about the potential for cloud computing with Matlab like computation.  Paul Gribble at GribbleLab has a post on high performance computing in the cloud where he details some of his computational requirements for arm simulations (24-36hr runs on whatever machines he can grab, 8-10 at a time max).

He also notes it would be cool if he could instead run many Amazon EC2 instances.  (And he can, see our earlier post).  But this clearly illustrates the problem – why should Paul have to beg borrow and steal computational resources when they already exist (and wouldn’t have to run overnight when the desktops aren’t otherwise in use, etc).

Paul’s commenters note that Amazon has limits to its scale, and 10k CPUs isn’t accessible anytime soon, but the principle is sound and we’re here to help Paul and others with similar needs get there simply and quickly.

If you search for cloud matlab you’ll find result #2 is the following question on Matlab’s help pages:

Can MATLAB take advantage of a cloud computing service… ?

This states that it is possible to run Matlab on Amazon EC2, and if you share name, phone, email, and more they’ll share this whitepaper with you detailing how to do it (it sounds complex but feasible).  Note that the paper is written for system administrators, not users.

Really, MathWorks – do you think the majority of your users actually want to setup their own Amazon EC2 instances, generate their own AMIs, manage their own backups at yet another location, and then purchase additional *expensive* Matlab licenses for the pleasure of all this?

To reiterate our raison d’etre – we think not.  We think professionals (such as ourselves) are annoyed whenever system administration, complexity, and cost stand in between us and our important algorithms, experiments, and computations.  We think professionals want a service like Monkey Analytics that delivers scalable computation in a box.  (We also think they want cloud features, not thick client software, but maybe that’s a generational thing).

We’re here, we’ve launched, and we believe in what we’ve built.  Come take a look and see if you don’t agree.  Let us know either way!