Comments on: More analytics in the cloud http://jtonedm.com/2008/11/25/more-analytics-in-the-cloud/ James Taylor on Everything Decision Management Mon, 14 May 2012 18:05:17 +0000 hourly 1 http://wordpress.org/?v=3.3.2 By: Abhimanyu http://jtonedm.com/2008/11/25/more-analytics-in-the-cloud/comment-page-1/#comment-19829 Abhimanyu Tue, 23 Mar 2010 21:23:51 +0000 http://jtonedm.com/?p=658#comment-19829 I agree. The term "Analytics" today can be divided into 2 different sub-domains (according to IDC) viz. QR (Query and Reporting) and Advanced Analytics. Terabyte scale analytics is really only possible for QR analytics. For advanced analytics (statistics, machine learning etc.) today's technology (SAS, SPSS, R, Matlab) hardly scale to a few thousand (hundreds at max). So really, advanced analytics is still being run in the order of < 1GB. I agree. The term “Analytics” today can be divided into 2 different sub-domains (according to IDC) viz. QR (Query and Reporting) and Advanced Analytics. Terabyte scale analytics is really only possible for QR analytics. For advanced analytics (statistics, machine learning etc.) today’s technology (SAS, SPSS, R, Matlab) hardly scale to a few thousand (hundreds at max). So really, advanced analytics is still being run in the order of < 1GB.

]]> By: James Taylor http://jtonedm.com/2008/11/25/more-analytics-in-the-cloud/comment-page-1/#comment-11652 James Taylor Fri, 05 Dec 2008 02:48:42 +0000 http://jtonedm.com/?p=658#comment-11652 Well of course that's a really good question. I think this works well for compute-heavy (rather than data-heavy) modeling and for executing models once they are built. Like you I suspect that pushing lots of data into the cloud will remain a problem for a while. Well of course that’s a really good question. I think this works well for compute-heavy (rather than data-heavy) modeling and for executing models once they are built. Like you I suspect that pushing lots of data into the cloud will remain a problem for a while.

]]> By: Alex Esterkin http://jtonedm.com/2008/11/25/more-analytics-in-the-cloud/comment-page-1/#comment-11650 Alex Esterkin Fri, 05 Dec 2008 02:44:57 +0000 http://jtonedm.com/?p=658#comment-11650 How much analytical data can be realistically processed in this way.   For example, will this work if I have 10 terabytes of data to store in an analytical database?  50 terabytes?   200 terabytes? How much analytical data can be realistically processed in this way.   For example, will this work if I have 10 terabytes of data to store in an analytical database?  50 terabytes?   200 terabytes?

]]> By: Michael Zeller http://jtonedm.com/2008/11/25/more-analytics-in-the-cloud/comment-page-1/#comment-11621 Michael Zeller Mon, 01 Dec 2008 19:12:29 +0000 http://jtonedm.com/?p=658#comment-11621 It is great to see more mathematical applications migrate to the cloud. This is one of the best opportunities where cloud computing can reduce cost and complexity of implementing computational efforts in HPC, large-scale simulations, and predictive analytics. As James mentioned, we at <a href="http://www.zementis.com" rel="nofollow">Zementis</a> launched the ADAPA predictive analytics decision engine on Amazon EC2 which allows users to deploy, integrate, and execute statistical scoring models, e.g., using algorithms like neural networks, support vector machine (SVM), decision tree, and various regression models. As the model exchange format, we leverage the <a href="http://adapasupport.zementis.com/2008/10/list-of-pmml-consumers-and-producers.html" rel="nofollow">Predictive Model Markup Language</a> (PMML) standard which is supported by commercial vendors like SPSS, SAS, IBM, Microstrategy, etc. as well as <a href="http://adapasupport.zementis.com/2008/02/how-can-i-export-pmml-code-from-r.html" rel="nofollow">open source tools like R</a> . With open standards and cloud computing platforms available, we hope to see more solutions emerge! It is great to see more mathematical applications migrate to the cloud. This is one of the best opportunities where cloud computing can reduce cost and complexity of implementing computational efforts in HPC, large-scale simulations, and predictive analytics.
As James mentioned, we at Zementis launched the ADAPA predictive analytics decision engine on Amazon EC2 which allows users to deploy, integrate, and execute statistical scoring models, e.g., using algorithms like neural networks, support vector machine (SVM), decision tree, and various regression models.
As the model exchange format, we leverage the Predictive Model Markup Language (PMML) standard which is supported by commercial vendors like SPSS, SAS, IBM, Microstrategy, etc. as well as open source tools like R .

With open standards and cloud computing platforms available, we hope to see more solutions emerge!

]]>