7 Replies Latest reply on Oct 26, 2011 3:29 PM by spilon

    Gartner Identifies the Top 10 Strategic Technologies for 2012

    wpurrer

      http://www.gartner.com/it/page.jsp?id=1826214

       

      a couple of them - osisoft related -  .. how is Osisoft playing this game ?

       

      Next-Generation Analytics. Analytics is growing along three key dimensions:

      1. From traditional offline analytics to in-line embedded analytics. This has been the focus for many efforts in the past and will continue to be an important focus for analytics.
      2. From analyzing historical data to explain what happened to analyzing historical and real-time data from multiple systems to simulate and predict the future.
      3. Over the next three years, analytics will mature along a third dimension, from structured and simple data analyzed by individuals to analysis of complex information of many types (text, video, etc…) from many systems supporting a collaborative decision process that brings multiple people together to analyze, brainstorm and make decisions.

      Analytics is also beginning to shift to the cloud and exploit cloud resources for high performance and grid computing.

       

      In 2011 and 2012, analytics will increasingly focus on decisions and collaboration. The new step is to provide simulation, prediction, optimization and other analytics, not simply information, to empower even more decision flexibility at the time and place of every business process action.

       

      Big Data. The size, complexity of formats and speed of delivery exceeds the capabilities of traditional data management technologies; it requires the use of new or exotic technologies simply to manage the volume alone. Many new technologies are emerging, with the potential to be disruptive (e.g., in-memory DBMS). Analytics has become a major driving application for data warehousing, with the use of MapReduce outside and inside the DBMS, and the use of self-service data marts. One major implication of big data is that in the future users will not be able to put all useful information into a single data warehouse. Logical data warehouses bringing together information from multiple sources as needed will replace the single data warehouse model. 

       

      In-Memory Computing. Gartner sees huge use of flash memory in consumer devices, entertainment equipment and other embedded IT systems. In addition, it offers a new layer of the memory hierarchy in servers that has key advantages — space, heat, performance and ruggedness among them. Besides delivering a new storage tier, the availability of large amounts of memory is driving new application models. In-memory applications platforms include in-memory analytics, event processing platforms, in-memory application servers, in-memory data management and in-memory messaging.

       

      Running existing applications in-memory or refactoring these applications to exploit in-memory approaches can result in improved transactional application performance and scalability, lower latency (less than one microsecond) application messaging, dramatically faster batch execution and faster response time in analytical applications. As cost and availability of memory intensive hardware platforms reach tipping points in 2012 and 2013, the in-memory approach will enter the mainstream.

        • Re: Gartner Identifies the Top 10 Strategic Technologies for 2012
          Ahmad Fattahi

          This is a very important point Wolfgang thanks for sharing it. On the highest level obviously PI System is made to handle large amounts of data, deliver, visualize, and analyze. Depending on the specific analytics you have in there may be a product either existing or planned for release to cover that (PE, PI ACE, PI for StreamInsight, Configured Analytics).

           

          Of course there are many more advanced analytics features that are not meant to be included in the PI System as well. There are two major questions:

          • Do you have any specific analytics/topic in mind?
          • Would you like to see them implemented as part of the PI System or want to see easier ways to integrate with more advanced analytical platforms (MATLAB, SAS, Mathematica, etc.)?

          I am sure you are aware of our white papers on how to integrate PI System with MATLAB and PI. I am personally very interested in this area. I would love to talk to you in vCampus Live! about that. I also have my presentation this year focusing on an aspect of this topic "Machine learning for prediction purposes on PI System data".

            • Re: Gartner Identifies the Top 10 Strategic Technologies for 2012
              wpurrer

              Hi Ahmad.

               

              In this area, for me there are two things important.

               

              At first, streamlining the process of making a trend and then further analyze it in a datamining tool.

               

              That it all what the Matrix plugin is about. We are now having the sixt or seventh Version.

               

              You select a trend in processbook, convert the data in for example hourly averages, fill up some missing data, filter some data, export it into a data analytical tool (for example weka) do some statistical analysis of the data and modify it (for example normalize it) then put into a software which creates a model. (Heuristiclab). This can be done in less then 5 minutes ( faster then an ignite presentation).

               

              As always, we have spend at list 15k achieve this goal. Why doesn't osisoft provide this .. integrated in it's products. (Please give some productmanagers some tasks where they have to really use their own product,... laurie one week in Lenzing ;-)  I have spent in the last couple of weeks a lot of time with pi and datamining, especally because I'm writing a Master Thesis in this area

               

              The second thing i forgot... some ideas about products that already exist which are really designed, for analyzing production process data

                • Re: Gartner Identifies the Top 10 Strategic Technologies for 2012
                  wpurrer

                  Ps it would really be nice to tell the pi server to have the last 2 month of data in working memory to speed up the displaying of the data.  (Would be 32 Gb in Our case) or maybe the last year ...

                    • Re: Gartner Identifies the Top 10 Strategic Technologies for 2012
                      Ahmad Fattahi

                      Thanks for the reply Wolfgang. I personally agree with you that streamlining the data mining capacities on PI Data is of extreme value to the PI System users in the future. The volume of data and also expectations are getting so high that we need new tools.

                       

                      Taking the extra step of integrating PI System with another platform (such as MATLAB or SAS) and writing the code is very powerful and may be the way to go for some organizations. However, for some others -maybe in your case- this is too much to ask. For smaller organizations especially, such out-of-the-box features would be very helpful.

                       

                      May I ask more details of what you want to do on your data? What data mining operations you have in mind?

                        • Re: Gartner Identifies the Top 10 Strategic Technologies for 2012
                          wpurrer

                          cause and effect analysis, time series modelling,...

                           

                          Even if some org, have the capablities .. it would be nice for all orgs. to have some out of the box. Where you don't have to take care about maintenance of the solution, integration problems,...

                           

                          maybe we have a chat during vcampus on this topic.

                            • Re: Gartner Identifies the Top 10 Strategic Technologies for 2012
                              Ahmad Fattahi

                              Sure let's plan on doing that. I am very interested in hearing your requirements and ideas.

                                • Re: Gartner Identifies the Top 10 Strategic Technologies for 2012

                                  @Wolfgang: thanks again for the important discussion and questions.

                                   

                                  You asked "how is OSIsoft playing in this game?"... well our strategy has been and continues to be to improve our PI Data Access mechanisms to play well in these scenarios. We definitely don't want to compete with Microsoft, SAP, IBM, etc. in these spaces. We are a real-time data and event *infrastructure*.

                                   

                                  We know there are some things we need to keep working on to fully achieve that (both performance- and functionality-wise), and I would be curious to hear your thoughts on the topic - i.e. what do you think is missing in terms of data access protocols, mechanisms, etc.

                                   

                                  While we can talk about this here, I see you will be at OSIsoft vCampus Live! this year and I look forward to discussing this when we meet in-person :)  In fact you might have noticed we have scheduled a Roundtable Discussion session around PI System Integration (Wednesday, 6pm)... this is obviously not the only opportunity we will have to discuss that, but hopefully other people who care about this will join the session too!

                                   

                                  Wolfgang Purrer

                                  Please give some productmanagers some tasks where they have to really use their own product...
                                  We definitely all use our products and try to steer them in the right direction for our users, but there's nothing like taking real-life scenarios from users like you to inspire ourselves from. As a matter of fact, it is our mandate to stay in touch with customers and partners, and make sure we address their day-to-day pain points as much as possible. That is how we try to communicate/demo our products most of the time, but obviously there are scenarios we don't address.

                                   

                                  And this is why I am particularly interested in speaking with you about all this, and see how you use (or wish to use) our products in conjunction with products like the Matrix Plugin, Weka, and the other data mining and production data analytic tools you referred to. I must admit I would definitely like to take you up on your offer (to Laurie) to spend a week in Lenzing... but in the meantime I'll contact you to see if we can talk about this between now and the vCampus Live! event.

                                   

                                  Let's make this masters thesis a blast!