Docstoc

Harry

Document Sample
Harry Powered By Docstoc
					As quick as you think!




                06       Paul Landucci
                         Stéphane Coez
Our Job
Applications deployed
through the WEB and
      Windows
Analysis and reports for
 developers & expert
         users
HARRYSoftware,
  Our Clients
  The company asset!
          HARRY Foundations
• Designed for business user
  requirements
• Powerful and complete vs.
  competitors
• One single interface to Relational and
  OLAP environments
• Consistent, fast and customized
  deployment
The HARRY Suite
    Version 7
      HARRYPilot, Version 7
 One single GUI interface to relational,
 multidimensional and vector DB:
• Accessing ORACLE, ESSBASE, MS-SQL
  Server, MS OLAP, HARRYVS… from your
  dictionary
• Design parameterized queries, through the
  same GUI interface and automatisms,
• Navigate within OLAP environments,
• Transform any query to an application page or
  broadcast it…
     HARRYPilot
To structure & display
     information
     HReport,
powerful & pertinent
   presentation
HARRYDataWEB
    • To design the
      application pages and
      dynamics,
      and run it on your
      Intranet,
    • To broadcast results
      and information,
    • To deploy according
      to your user domains,
      groups and profiles.
            A 4 step process
• Step 1: Create your report at
  the          detail level

• Step 2: Replace pertinent
  values
          by parameters

• Step 3: Insert and design within
  a
           WEB application page
                   HARRYDataWEB
                 Interactive reporting,
                     Ad hoc queries,
                   HARRYNetServer
               Mutidimensional navigation

                   HARRYDashboard
 HARRY
Manager,   HCube                      HReport
  HAS
                    HARRYPilot


                    Dictionary
HARRYVectorServe
      rs
    Very fast datamarts:
        HARRY 2: IBM Z90
       HARRY-X: AIX, SUN, HP
    HARRY-NT: WINDOWS XP & 2003
           HARRY-LINUX
Vector Architecture
          Adapted to:
          • Marketing,
          • Sales, logistics,
          • WEB logging
            analysis,
          • …
            Any large
            database
    The VectorServers many
           benefits:
• Extremely fast on complex queries, scan,
  calculations,
• Unlimited number of columns or
  dimensions,
• Boolean fields and bitmap indexing,
• Dynamical functions and dimensions,
  – Ranging or intervals
  – IF(CONDITION;TRUE;FALSE),
  – Typologies or Segment
         Selection: Where and Count

                        Active selection                                Vector table
                            bitmap
                              0
                              1
                              0       Block 1                       Data
 Partition                    0
  bitmap                      0
                              1




                                                                                        Unit Price
                                                              Product


                                                                             Quantity
                              1




                                                       ID #




                                                                                                     Etc.
                                                Date
     1                        0       Block 2
     1                        1
     1
     0
                              1                                                                             …
                              0
     0                        1
     0                        0       Block 3
                              0
                              0
                              0
                              0
                              0       Block 4
                              0
    VectorServer is           0
 working on partition         0
                              0
 and blocks. Block is
the minimum read unit             As quick as you think!
     Vector index

                                                              Selection
                                                               bitmap

                                                                 0
              Addresses in                                       0
                                                                 1
 Code         vector table                                       0
                                                                 1
                                                                 0
         74   5,15,194,2056,…        Bitmap creation             0
         75   2,3,12,16,17,18,27,…                               0
                                                                 0
         76   1,8,10,…                                           0
                                                                 0
 …                                                               0
                                                                 0
                                                                 0
                                                                 0
 Index allows:                                                  0
                                                                 1
                                                                 0
      To create a selection bitmap from a                       0
                                                                 0
       matrix describing each value with                         0
                                                                 1
       addresse for a variable in table                          0
                                                                 0
      Without file reading                                      0
                                                                 0



                                     As quick as you think!
  APL/M conversion



 APL/M Version 10 & 11…

    Availability on WINDOWS, UNIX, LINUX
    Version 7.1 de HVS, released




 Overpasses the 4 Giga Bytes file limit

 UNIX, LINUX et WINDOWS files directly compatible




                       As quick as you think!
Nouveau look & feel XP
pour la console et le monitoring




                  As quick as you think!
Performances: 6.5 M Rec, 315 cols




              As quick as you think!
  HARRYVectorServer, next…



 64 bit full version

 Administration module as a ‘console’

 Stored relational views ‘join-like’

 Parallelization, multi-processing & clustering
    Vector loaded in // on several instances
    Dispatch one query on several servers
    Load balancing
    Dynamical profile session update




                         As quick as you think!
The fashion show room


      JAVA…!
A alternative new architecture

                     HARRYDataWEB
                   Interactive reporting,
                       Ad hoc queries,
                                    WEB

                 Mutidimensional navigation

    HCube             Application (Java)       HReport

              HARRYPilot & Dictionary
                  HARRY JAVA Classes


                     HARRY Workspace
                          VectorServer



                     Auxiliary processors

                             HARRY
                             Vector
            Oracle                              XX DB
                      As quick as you think!
  The HNS JAVA interface




 To give to developers a total and accurate control on
  HARRYVS at the appropriate level for their application.

      Access to every command or function and parameters of
       HARRYVectorServer.

       Example: HARRY « SEL:FAMILY=702 »
         Loads the selection bitmap in WS
         Returns the count to application


      Access to every HARRYVectorServer features or facility



                           As quick as you think!
  Auxiliary processors


  Goals : To create a dynamic link between an external server
  (Oracle, SQLServer, DB2/U, etc.) and the
  HARRYVectorServer WS through this server native
  middleware.
   Runs a query and sends the result to WS.
   Notice: Query is only checked by the remote remote server.


 Example: SQLORACLE

 Syntax:   SQLORACLE:SQL_QUERY
 Example : SQLORACLE:SELECT NAME,FISRT,SALARY
                      FROM PERSO WHERE
                      DIV=‘MFG'

                       As quick as you think!
       Example: a two server query

                          Select the customers who have today receivable
               code                  for more than 1,500 Euros
              & count
  Client
Application                       1 – ORACLE Query


                           « SQLORACLE:SELECT CUSTID,COUNT
                HARRY_EXECCust_ID + return
                                   LIST                            Workspace
                             list   code                              HVS
                          WHERE account >= 1500»




 ORACLE


        SELECT CUSTID,COUNT
                LIST
        WHERE account >= 1500
                                    As quick as you think!
       Example: a two server query

                          Select the customers who have today receivable
     Selected customers              for more than 1,500 Euros
           number
  Client
Application
                          2 – Key selection within vector table
                                           Cust_Id
                                           vector

                    HARRY_EXEC « CSEL:LIST1 IN Cust_ID
                                                                          Workspace
                                                                             HVS
                              Customer                       S1
                                 list




 ORACLE


                                          Load vector
                                                             HARRYVectorServer

                                    As quick as you think!
  HARRYSoftware



        www.harrysoftware.com

 Stéphane Coez, R&D
  scoez@harrysoftware.com

 Paul Landucci, CEO
  planducci@harrysoftware.com




                    As quick as you think!

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:8
posted:11/9/2011
language:English
pages:30