board technology overview

24
BOARD Technology Overview

Upload: others

Post on 15-Oct-2021

5 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: BOARD Technology Overview

BOARD Technology Overview

Page 2: BOARD Technology Overview

2 3Contents Contents

CONTENTS1 PLATFORM OVERVIEW...............................................................................................................6 1.1 Analyze, predict, plan and simulate in a single platform...................................6 1.2 Self-service analysis meets data governance: a two-way process..................8 1.3 Product architecture..................................................................................................10

2 DATA INTEGRATION..................................................................................................................14 2.1 Data federation...........................................................................................................16 2.2 BOARD Connectors...................................................................................................17 2.3 Multidimensional data sources...............................................................................17

2.4 BOARD Connector for SAP......................................................................................18

2.5 Essential ETL functionalities.....................................................................................20

2.6 Data Fast Track............................................................................................................21

3 BOARD ENGINE.........................................................................................................................22 3.1 Databases...................................................................................................................22

3.1.1 The HBMP technology: the new in-memory frontier........................23

3.1.1.1 HBMPbenefits:speed,concurrencyanddatavolumes...24

3.1.1.2 HBMP:thearchitecturaladvantage......................................24

3.2 Business logic.............................................................................................................25

3.2.1 Definingthebusinesslogic......................................................................25

3.2.2 Predictiveandadvancedanalytics(BEAM).........................................26

3.3 Application..................................................................................................................28

3.3.1 Thetoolkit:thefastestwaytobuildanalyticapplications................28

3.3.2 Storytelling...................................................................................................29

3.3.3 In-contextCollaboration..........................................................................30

3.3.4 Cognitive technologies.............................................................................25

4 CLIENTS........................................................................................................................................33 4.1 BOARD Web...............................................................................................................33 4.2 BOARD Mobile...........................................................................................................34

4.3 BOARDOfficeAdd-Ins.............................................................................................35

4.4 BOARD Windows Client...........................................................................................36

5 SECURITY......................................................................................................................................37

5.1 Multi-tier security model..........................................................................................37

5.2 Authentication methods..........................................................................................38

5.3 Role based authorization........................................................................................39

5.4 Application authorization.........................................................................................39

5.5 Data access.................................................................................................................39

5.6 Advanced security management...........................................................................40

6 INFRASTRUCTURE.....................................................................................................................41

6.1 Cloud or on-premises..............................................................................................41

6.2 In-memory cluster: horizontal scalability.............................................................42

6.2.1 HowtheBOARDClusterworks..............................................................44

6.2.2 Adaptiveloadbalancing..........................................................................44

6.3 Interoperability and portal integration.................................................................45

6.3.1 Webmash-upandportalintegration...................................................45

6.3.2 SharePoint integration.............................................................................45

6.3.3 BOARDwebservices.................................................................................46

Page 3: BOARD Technology Overview

5Platform Overview

Analyze, predict, plan and simulate in a single platform BOARD is a unified platform for decision-making that combines business intelligence,advanced analytics, planning and simulation capabilities into a single environment. These capabilities work in concert to provide the fastest and easiest way to build highly customized analytical and planning applications.The all-in-one approach means that BOARD provides all the capabilities needed to build, manage and maintain analytic solutions for:

BOARD All-in-One approach

OVERVIEW1.BOARDmerges data analysis and simulation into the planning, workflow and controllingcycles of organizations; transforming data into insights, insights into predictions, predictions into plans and plans into traceable and monitored actions at a strategic, financial andoperational level.

All solutions created in BOARD share the following elements:

• Single logical view of data

• Single and shared metadata across all types of solutions (BI, Planning, Simulation, Analytics)

• Single user interface

• Single security

• Single administrative environment

• Single visual modeling environment

Integratingtheseelementsintoaunifiedplatformcreatestremendoustechnicalefficiencies,whichBOARDleveragestoprovideanunrivaledcombinationofspeed,easeofuse,flexibilityand minimal cost of ownership. This allows BOARD customers to build and maintain their own bespoke solutions at a fraction of the time and cost required for traditional purpose-built applications.

Range of capabilities needed for delivering BOARD “All-in-One” approach

1.1

• Reporting and interactive dashboarding

• Ad-hoc analysis and data discovery

• Predictive analytics and simulation

• Budgeting, planning and forecasting

• Profitability modeling and financial consolidation

• Scorecarding and strategy management

Unified Metadata / Data Federation

ScorecardingDashboardReportingAnalysis

Budgeting, Planning & Forecasting

Strategy ManagementProfitability Analysis Simulation

Financial Consolidation

Forecasting

ClusteringStatistical Functions

ERP, CRM BIG DATA,NoSQL DATABASESLEGACY SYSTEM

RELATIONAL DATABASES,

DATA WAREHOUSE

OLAP CUBESCLOUD SOURCES

EXCEL, CSV, TXT

BUSINESS INTELLIGENCE

PERFORMANCE MANAGEMENT

PREDICTIVEANALYTICS

Self ServiceAnalysis

PredictiveAnalysis

Simulationwhat if

Planning andControlling

Reporting Dashboarding

Scorecards

Page 4: BOARD Technology Overview

6 7Platform Overview Platform Overview

Self-service analysis meets data governance: a two-way process Self-servicebusiness intelligenceprovidesvaluablebenefits intermsofagility,velocityanduser independence. However, if not combined with a proper data governance process, it introduces many long-lasting problems, including: an environment of proliferating BI silos, inconsistentbusinessdefinitions,nodatalineageandnosingleversionofthetruth.

BOARDoffers an innovative approach togoverneddatadiscovery, empoweringbusinessusers with full self-service capabilities while allowing the organization to maintain full control of data security, integrity and the degree of self-reliance.

From one side, business users can customize any number of centrally-authored reports and visualizations and then save them to a personal workspace. This way, users can create their own personal perspective on enterprise reports, underpinned by a validated and centralized system of record.

From the center to the Edges:Users create their own customized version of corporate reporting

From the Edges to the center:Users created their own reports that can be validate and included into corporate reporting/environment

Fromtheotherside,userscanimporttheirowndataintoasandboxenvironmentanduseit for independent analysis. If the insights provided by the new data are relevant to others, the analysis and the underlying data can be easily posted to a common repository and, once validated by an administrator, promoted and shared across the organization.

This two way-process fully supports the self-service needs of the organization, in a trusted and validated environment, backed by the capability to trace the data lineage back to its source and to control how the data was manipulated before being analyzed.

1.2

Data

Data

Data

Data

Data

Data

Datavalidation

Datavalidation

Data

Data

Data

Data

Data

Data

Datavalidation

Datavalidation

Page 5: BOARD Technology Overview

8 9Platform Overview Platform Overview

BOARD’s functional architecture includes the following elements:

Data integrationFrom relational databases to cloud applications, from big data to cloud data stores, it does notmatterwhichdatayouaccess,norwhereitisstored.BOARDoffersasuiteofconnectivitytools to streamline access to virtually any data source.

BOARD EngineThe BOARD Engine encompasses all the functionality required to model, build and manage solutions on the platform and is based on three building blocks: databases, business logic and applications.

• Databases: Databasesprovidethecapabilitytoaccess,normalize,federateandmanagedata.

• Business logic:Acomprehensivesetoffunctionalitiesandbusinessrulestoimplement,runandmanageanalysis,planning,budgeting,forecasting,consolidation,allocationandsimulationlogic–seamlesslylinkingdatawithbusinessprocesses.

• Applications:Theend-userenvironment thatallowsusers toview, interactwithandanalyzedata.Non-technical users can become developers, quickly building applications withoutcoding,leveragingallthecapabilitiesofferedbythedatabasesandthebusinesslogic.

ClientsBOARDensuresaconsistentexperienceacrossweb,mobileanddesktopenvironmentssodevelopers can build a single application and deploy it to users on variety of devices. BOARD applications can be accessed through the following clients:

• BOARD Web (HTML5)

• BOARD Mobile

• BOARD Office Add-Ins

• BOARD Windows Client

Infrastructure & SecurityBOARD offers an enterprise-grade infrastructure to support the largest worldwidedeployments.Thisincludesstrongandflexiblemulti-tiersecurity,custom-tailoredcloudandon-premises networking, as well as advanced in-memory clustering capabilities.

Product architecture

The BOARD platform ensures all the scalability, security and data-governance capabilities that IT needs to deploy and support enterprise-class business intelligence and planning projects.BOARD’smulti-tierarchitectureoffers:

• Scalability BOARDbeendesignedtoensurehigh levelsofscalability in termsofconcurrentuserrequests,datavolumesandthescope/complexityofapplications.Its integrated in-memory clustering technology supports large (thousands of users)globaldeploymentsandavoidsproblemswithgeographiclatency.

• PerformanceBOARD offers outstanding performance, not only in terms of data aggregation andpresentation,butalsoinhandlinglargeandcomplexplanningprocesseswiththousandsofusersperformingconcurrentdata-entry.This ismadepossible by the combinationof BOARD’s hybrid in-memory technology(HBMP) with a powerful and efficient multidimensional database that includessophisticatedsparsedatamanagementtechnology.

• Consistent experience AnyapplicationbuiltinBOARDcanbeaccessedbymultipleclients(webbrowser,BOARDWindows Client, mobile apps, Office add-in) providing a consistent user experienceacrossalldevices.

• Robust securitySecurity is pervasive across the whole platform, from authentication to cell levelauthorization.BOARDoffersarobust,comprehensiveandeasytomanageenterprise-classsecuritysystem.

• 24 x 7 availabilityTheflexibilityofBOARD’sdatamodelallowsuserstomakechangesinrealtimeandtoperformonlinedataupdates,whichmakesitpossibletoprovideorganizationswith24x7availability.

1.3

Page 6: BOARD Technology Overview

10 11Platform Overview Platform Overview

WINDOWS CLIENT

HTML5 WEB CLIENT

OFFICE ADD-IN

TABLET MOBILE

PORTAL MASHUP

CAPSULE A

CAPSULE B

CAPSULE C

ALERTING

BROADCASTING

WORKFLOW

ALLOCATION CONSOLIDATION,

SIMULATION

DATAFLOW

CUBES

DATAENTRY

HIERARCHIES

HETEROGENEOUS DATA

DIMENSIONS

UNIFIED METADATA

1101110....abcdef...

APPLICATIONSBOARD DATABASES BUSINESS LOGIC

BOARD ENGINE

ERP, CRM LEGA

BIG DATA,NoSQL DATABASES

CY SYSTEM

CLOUDSOURCES

EXCEL, CSV, TXT

OLAP CUBES

RELATIONAL DATABASES,

DATA WAREHOUSE

DATA SOURCES CLIENTS

BOARD: Logical Architecture Schema

Page 7: BOARD Technology Overview

12 13Data Integration Data Integration

STORED CUBES VIRTUAL CUBES

FRONT END

HBMPDUAL ENGINE

Data Upload

Data Upload

Data Upload

Read and Write

Read

OLE

DB

O D

ATA

OLE

DB

OD

BO

Read

O D

ATA

RDBMS CLOUD CLOUDRDBMS

OD

BO

OLAP SERVER OLAP SERVERS

MOBILE TABLETWINDOWS CLIENT WEB CLIENT OFFICE ADD-IN MASHUPPORTAL

Throughitsphysicalandlogicalmultidimensionaldatabase,BOARDoffersasinglemetadatagateway that makes it possible to use data as if it were native to BOARD. Data is normalized and organized within BOARD, allowing end-users to read, write and update the data, regardless of the source.

Behind the scenes, the administrators have the freedom to decide which data to store within BOARD’s multidimensional database (MOLAP) and which to leave in place and manage through BOARD data federation capabilities (ROLAP). The BOARD DUAL Engine will automaticallymanagethetwodifferentoptions,providinguserswithasingleandseamlessvision of the data.

BOARD also supports write back to both its multidimensional database as well as to relational data sources, making it possible to integrate performance management processes with enterprise applications (e.g. storing forecasting/planning data directly within the ERP).

Access to these data sources is provided through the following technologies:

• Native pre-built data connectors

• Open Database Connectivity (ODBC) and OLE DB to connect with rela-tional databases

• ODATA, JSON and REST to connect with cloud based sources

• OLE DB for OLAP (ODBO) to connect with multidimensional databases

• BOARD connector for SAP

• Native text file reader

DATA INTEGRATION2.

While maintaining data within a single enterprise data warehouse can provide substantial benefitstoanorganization,businesscomplexityoftenmakesthistoodemanding,difficultoreven impossible. In practice, large volumes of data are frequently stored in a wide variety of sourcesandplatforms,bothinternalandexternaltotheenterprise.

The BOARD platform provides the capability to connect and integrate data across virtually any type of data source, including:

• Relational databases and data warehouses (e.g. Oracle, SQL Server)

• Enterprise applications (e.g. SAP ERP)

• Cloud based sources (e.g. Salesforce, Google Analytics)

• Column oriented databases (e.g. Teradata, SAP HANA)

• Big Data/NoSQL databases (e.g. Hadoop, Impala)

• Multidimensional sources (e.g. SAP BW)

• Web services

• Excel, CSV and TXT files

Page 8: BOARD Technology Overview

14 15Data Integration Data Integration

Data federation

Data federation is the process of gathering data from disparate data sources without transferring the original data. Rather than combine all of the source data into a new database, physical data is left in place and metadata - data that describes the data and its location - is utilized to create a virtual database.

BOARD’s data federation capabilities (ROLAP engine) provide the ability to read and write data stored in a RDBMS as if it were stored in the BOARD multidimensional database.More precisely, metadata is stored in BOARD’s virtual cubes; whenever a user views a report basedonthesecubes,aconnectiontotheexternalrelationaldatabase isestablishedanddata is retrieved from the relational table through a SQL query.

Anytime a user runs a request that accesses multiple relational sources, BOARD automatically:Generates a series of queries for each database needed by the report/analysis, connects to thecorrespondingdatasourcesforexecution,feedstheBOARDvirtualcubeswiththefinalresultsfromeachofthequeriesandtransparentlymakesthefinalmergedresultsavailableinthe report/analysis

From an architectural standpoint, this capability allows developers:TobuildextremelyefficientBIandCPMapplicationsbyeliminatingtheneedtouploaddataatthe lowest level of detail into BOARD databases, without compromising the ability to access this data. In fact, data can be physically stored at the most suitable aggregation level for each analytic model, thus reducing data volumes and time-consuming upload processes while granting the ability to access the lowest level of detail through virtual cubes whenever needed.To create near real-time analytic models by making RDBMS data directly available in the BOARD MOLAP environment.

BOARD Connectors

BOARDoffersacomprehensivesetofpre-builtdedicateddataconnectorsthatminimizethetimeandeffortrequiredtoaccessanduseyourdata.From relational databases to cloud applications, from big data to cloud data stores, it does not matter which data you tap, nor where it is stored.

2.22.1

ODBO is a Microsoft specification and an industry standard for multidimensional dataprocessingspecificallydesignedtoaccessmultidimensionaldatasources.XMLAisessentiallya newer standard interface. Both of these technologies operate using the Multidimensional Expressions(MDX)languagetoquerycubes.

• Microsoft Analysis Service

• Oracle/ Hyperion Essbase

• SAP BW

• IBM Cognos TM1

Multidimensional data sources

BOARD supports OLE DB for OLAP (ODBO) and XML for Analysis (XMLA), providingcustomers with a standard method of accessing the most popular multidimensional engines, such as:

2.3

Big Data

Amazon EMR HiveApache CassandraApache Hadoop HiveCloudera CDH HiveCloudera ImpalaDataStaxEnterpriseHortonworks HiveIBM BigInsights HiveMapR HiveMongoDBPivotal HAWQPivotal HD HiveSpark SQL

Relational and Analytics

Amazon RedshiftIBM DB2IBMInformixMicrosoft SQL ServerMySQL CommunityMySQL EnterpriseOracle DatabasePivotal GreenplumPostgreSQLProgress® OpenEdge®SAP Sybase ASESAP Sybase IQTeradata

SAAS/Cloud

Financial ForceForce.com ApplicationsGoogle AnalyticsHubspotMarketoMicrosoft Dynamics CRMMicrosoft SQL AzureOracle EloquaOracle Sales CloudOracle Service CloudProgress® Rollbase®Salesforce.comServiceMAXSugarCRMVeeva CRM

Page 9: BOARD Technology Overview

16 17Data Integration Data Integration

BOARD Connector for SAP

Harnessing data from SAP ERP is a notoriously difficult and sometimes a very expensiveactivity.BOARDoffersadedicatedconnectorforSAPenvironmentstostreamlinethisprocess.Easilyinstallableandconfigurable,theBOARDSAPConnectorallowsmappingSAPdataintoBOARDdatabaseswithoutwritinganABAPprogramfordataextraction.

Auser-friendlygraphicalinterfacemakesiteasytoimplementdataflowstoextractandloaddata into BOARD in a safe and secure manner, all with high performance and stability.

The BOARD SAP connector interoperates with all varieties of SAP interfaces by means of 8 specialized components.

Tables

This method allows you to extract mass data directly from SAP tables. To optimize the network load, only those columns that have been explicitly selected are extracted

BAPIs and function modules

Using the Xtract BAPI component, you can call BAPIs or RFC Function Modules directly from BOARD Data Readers

OHS (Open Hub Services)

Connect with SAP OHS. This component allows automatic extraction processes to be created, including to different SAP objects (not only cubes, but InfoObjects, ODS ob-jects, texts, etc.).

Reports

Use existing ABAP reports and programs as data sources within the BOARD Data Reader, making it possible to reuse existing logic. This technique is often used in the area of FI/CO reporting

Queries

This component makes SAP queries directly available in BOARD

BW-Cubes & BW-Queries

This component extracts datasets from SAP BW InfoCubes. Dimensions, key figures and attributes can be transferred by a simple drag and drop into the query output

Hierarchies

This method allows users to extract any hierarchy type from SAP BW. If required, hierarchy texts can be extracted as well.

DeltaQ

The DeltaQ component uses the SAP BW Extractor API (the functionality that SAP BW relies on to get its data from the live system), making information immediately available – without the need to understand and work on the underlying table structure

2.4 How it worksTheBOARDConnectorforSAPisamiddlewarecomponenttoextractdatafromSAPsystemsand load it into BOARD. It has two components: Designer and Server. The Designer allows a systemadministratortodesignandconfiguretheextractionsofSAPobjectssuchastables,views,queries,BWInfoCubes,BAPIfunctionmodulesetc.TheServerSAPAPIs(SAPCertified)textextractsdatafromSAPandpassesittoBOARD.

Morespecifically,theprocesscanbesplitintothesesteps:

• Extractions creation ThesystemadministratorusestheBOARDConnectorDesignertodefineacatalogofSAPobjects(tables,queries,ABAPprograms)thatcanbeexposedtoBOARDintheXtractslibrary.TheBOARDConnectorconnectstoSAPthroughtheRFCs,thestandardSAPAPIs.

• Data Reader configuration TheSystemadministratormapstheavailableextractorstoBOARDdatabases.BOARDDataReaderconnectstoBOARDConnectorthroughWebServices.

• Extractor Execution BOARDSAPConnectersendsarequesttoSAPandpassestheresultsettoBOARD.

Page 10: BOARD Technology Overview

18 19Data Integration Data Integration

Essential ETL functionalities

WhileBOARD isnotadedicatedETLproduct, itdoesoffercapabilities fordiversesourcesystems to directly feed BOARD databases without the need for intermediate data staging layers.This is unique when compared to other BI products, which typically require the source data tobecleansedandorganizedintoeitherastarorsnowflakeschema,whichiscostly,rigidand time consuming.

The key ETL capabilities provided by BOARD through its ETL and Data Reader features are:

• Validation rules definingsimpleorcompositevalidationformulas/criteriasuchas:

a.ignorethoserecordswhereacertainkeyorvalueisnotpresentb.ignoreoracceptrecordswherecertainfieldscontainagivenvalue(orthresholdorsetofvalues)

• Normalization and conversiona. Translatingcodedvalues(e.g.sourcesystemcontains1formaleand2forfemalebutwewanttouseMformaleandFforfemale)

b. Encodingfreeformvalues(e.g.mappingDEtoGermany,FRtoFrance)c. Standardizingcodesfrommultiplesources(e.g.onesourcesystemuses01foragivenproductandanothersystemusesABC)

d. Derivinganewcalculatedvalue(e.g.sales_amount=qty*unit_price)e. Definingdefaultvaluesfornullfields(e.g.ifcurrencyisnullthen“Euro”)f. Standardizingdateformats(e.g.onesourcesystemcontainsdatesinaformatofyyyy-mm-ddandanotheroneindd-mm-yyyy)

g. Uselogicaloperatorsforexpressionsandtransformationformulas

• Independence from the underlying datasourceBoardETLrulesaredefinedinanindependentenvironmentfromtheSQLusedtoquerytherelationaldatabasesofthedifferentdatasources,e.g.transformadateformatfromaMSSQLsourceandfromanOraclesourceusingasingleformulathatdoesn’tdependontheT-SQLorPL-SQLsyntax.

2.5 Data Fast Track

The Data Fast Track feature is for self-service data modelling, mash-ups and uploading. With Data Fast Track, business users can quickly create compelling visual analyses and analytical applications from raw data in just a few clicks, without any assistance from IT.

UsersconnecttoadatasetandBOARDwillautomaticallybuildadatamodelthatreflectstherelationships inherent in the source data. Users can then enrich the data model by blending additionaldatasetsorbycreatingderivedfieldsinanExcel-likeenvironment.

Oncethedatamodelisfinalized,userscanimmediatelycreatetheiranalysesusingBOARD’sdrag-and-drop environment.

• Allows business users to create analysis from raw data, injustafewclicks.

• Smart data blending fromdifferentsources.Automaticallymergediversedatasourcesintoasingledatamodel.

• Automatic discovery and classification of data types (measures, dimen-sions,hierarchiesandentities).

• Self-service data manipulation. Aspreadsheet-likeenvironmentallowsusertocreatenewmeasureswithformulas,tocutandpastedataandtoeditmeta-datamappings.

2.6

Page 11: BOARD Technology Overview

20 21BOARD Engine BOARD Engine

BOARD ENGINE3.The BOARD Engine encompasses all the functionality required to model, build and manage solutions on the platform and is based on three building blocks: databases, business logic and applications.

The BOARD Engine executes all aggregations, calculations, selections, procedures, dataimport processes and any other interaction involving the BOARD multidimensional database. It also handles incoming connections from BOARD users.

The BOARD Engine also performs user authentication, applies security restrictions or privileges andthendispatchesuserrequestsasseparateexecutionthreadstocarryouttheanalyticalprocessing.

Communication between the clients and the engine use a proprietary protocol named ROAR (RemoteObjectAccess&Replication)whichprovides extremely efficient communicationsandoffershighperformanceoverlowbandwidthconnections.

Databases

BOARD provides a single view of organizational data through its combination of physical and logical multidimensional databases. BOARD acts as a metadata gateway that allows all source data to be used as if it was native to BOARD. Data is normalized and organized in BOARD data models, allowing users to perform reads, writes and updates regardless of data sources.

Behind the scenes, administrators have the freedom to decide which data to store into BOARD’s multidimensional databases (MOLAP), and which to leave in place and manage through BOARD data federation capabilities (ROLAP). BOARD’s DUAL Engine automatically managesthetwodifferentoptions,providinguserswithasingleandseamlessvisionofthedata.

This model also applies to data-entry processes; BOARD supports write-back not only on its data model, but also directly to relational data sources.

• Databases Databasesprovidethecapabilitytoaccess,normalize,federateandmanagedata

• Business logicAcomprehensivesetoffunctionalitiesandbusinessrulestoimplement,runandmanageanalysis,planning,budgeting,forecasting,consolidation,allocationandsimulationlogic–seamlesslylinkingdatawithbusinessprocesses.

• ApplicationsTheBOARDend-userenvironmentthatallowsuserstoview,interactwithandanalyzedata.Developerscanquicklybuildandrunapplicationswithoutcoding,leveragingallthecapabilitiesofferedbythedatabasesandthebusinesslogic.

3.1

ThesethreedifferentwaysofconfiguringtheBOARDin-memoryenvironmentallowgreatflexibilityinchoosingthemodelthatbestsuitsthesolutionrequirementsandtomakeoptimaltradeoffsbetweenscalability,priceandperformance.

3.1.1 The HBMP technology: the new in-memory frontier

At its core, the BOARD platform is based on a revolutionary proprietary in-memory data management technology named HBMP.HBMP technology has been conceived with a precise and extremely ambitious goal: toexploit the full potential of in-memory computing (IMC) in terms of better performance,while improving theflexibility, scalabilityandability toeffectively supportdecision-makingprocesses that have fueled international success for BOARD.The result is a revolutionary solution that overcomes the limitations of standard BI in-memory technology, combining pure performance with the unique capability to support and digitalize simulation and planning processes across the enterprise.

Thetechnology’sname,HybridBitwiseMemoryPattern,reflectsitstwomaindifferentiatingcharacteristics: the Bitwise Memory Pattern and its unique Hybrid approach.

Bitwise Memory PatternThe term “Bitwise Memory Pattern” describes BOARD’s ability to store and map physical data through unique bits, or rather to “bitmap” them by means of an innovative proprietary algorithm. In other words, BOARD maps the compressed multidimensional data structure (Pattern) at bit level (Bitwise) into the RAM (Memory). The new algorithm brings enormous benefits,notonlyintermsofperformancebutalsointermsofdatacompression,parallelismand high-user concurrency. Unlike most in-memory solutions, the algorithm ensures full bi-directionality,whichsupportswrite-backand“onthefly”changesofthedatastructure.

Hybrid approachThehybridnatureprovidestheabilitytomanagedatainthreedifferentways:

• Full in memory: Allprocessingisperformedagainstdataheldin-memory • On-demand in-memory: Determinewhichsubsetofdatatousein-memoryandwhichtostoreondisk.E.gin-memorymetadataforread-onlyoperations,dynamicuploadofphysicaldataforwrite-back

• Hybrid in-memory: Usethedataindicesandmappingin-memorybutstoretherestofthedataondisk

Page 12: BOARD Technology Overview

22 23BOARD Engine BOARD Engine

3.1.1.1 HBMP benefits: speed, concurrency and data volumes

The HBMP technology is not limited to just storing data in-memory to reduce disk I/O, it also provides huge improvements in parallelism, data compression and user concurrency. In functional terms, this allows users to perform analysis and build simulation models on large volumes of highly granular operational data.

Parallel Multicore ProcessingHBMP technology parallelizes BOARD processing to fully leverage multi-core architectures. SpreadingthecalculationloadacrossCPUcoresenablesmoreeffectiveworkloadmanagementand a better handling of concurrency.

Data CompressionThe HBMP technology makes it possible to map the multidimensional structure of the data in memory at bit level. This compression scheme considerably reduces the space required to store and manage data, making the “weight” of BOARD databases similar to the most advanced read-only products.

ConcurrencyHBMP technology supports high levels of user concurrency, unlike traditional in-memory techniques that struggle with or require massive amounts of RAM to support large numbers of concurrent users.

3.1.1.2 HBMP: the architectural advantage

Three fundamental architectural factors that make it superior to other in-memory solutions in the market.

Bi-directional algorithmThemathematicalmodelbywhich thedata is indexedandmapped in-memorysupportsboth read and write operations and can distribute data changes along hierarchies.

In-memory metadataHBMP technology also performs the meta-data processing in-memory, not just data management like with other products. This allows the database structures and mappings tobechanged“onthefly”andtoimmediatelyseethechangesreflectedintheapplicationswithout having to reload the physical data. The combination of the in-memory metadata processing and the bi-directional algorithm is where HBMP technology derives its hybrid nature,i.e.theabilitytoadministerthemapping/indexingofdatain-memoryanddecidingwhether to physically store it in RAM or on disk.

Single server instanceHBMP technology supports concurrent users sharing a unique server instance. This minimizes theimpactofhighconcurrencyonmemoryresourcesandisdramaticallydifferentthanotherin-memory technologies, which are heavily impacted by concurrent user requests

Business logic

BOARD provides a comprehensive set of functionalities to implement the business logic that organizations need to support their analytic and planning processes.

Such processes include driver-based allocations, eliminations, conversions, consolidations, initializations, top down and bottom-up reconciliations and multidimensional calculations More generally, this encompasses all the necessary operations to implement company-wide planning, forecasting and simulation applications and includes functionality likeworkflow,scenario management and versioning.

3.2.1 Defining the business logic

Thebusinesslogiciscapturedwithinseveraldifferentareasoftheplatform,dependingonthedesired functionality. The most common are:

Regardlessofwherethelogiciscaptured,itisalwaysdefinedusingfunctionalbusinessterms,not code, within an agile and interactive modeling environment (the “toolkit”). This allows the business users themselves to quickly build customized solutions that perfectly fit theorganization’s requirements. The agile nature of the “toolkit” means that solutions can also be quickly adapted to ever changing business needs.

It’stemptingtothinkthatsincethelogiciseasytodefineanddoesn’trequirecoding,thenitmaylacktheflexibilitytosupportcomplex,real-worldusecases.Infact,theoppositeistrue.BOARDusersdefinethebusinesslogicusinganintuitive,parameter-basedapproach.Thisisnot only easier to learn but since the business logic is tightly coupled with the database in an integrated system (i.e. “data-aware”), users can make fundamental changes to the structure and contents of the underlying database and have these changes automatically propagated into the business logic.

Compare this with a traditional, code-based solution where the manual maintenance required to keep the hard-coded business logic in sync with the underlying data severely limits the scopeandcomplexityofthesolutionsthatcanbeprovided.

Whilethebusinesslogicisdefinedusingfunctionalbusinessterms,thelogicaloperationsthatthe underlying engine supports are highly advanced.

3.2

• Procedures

• Report and visualization objects within application screens

• BOARD’s predictive engine (BEAM)

Page 13: BOARD Technology Overview

24 25BOARD Engine BOARD Engine

3.2.2 Predictive and advanced analytics (BEAM)

You can embed advanced analytical and predictive capabilities into your business processes with the agility necessary to operate at the ultra-rapid pace of today’s business. And do it in a simple way, allowing users to take the lead, without having to depend on an armyofexpertslikewiththetraditionaldata-miningandanalyticsproducts.

Someexamples:

• User-defined formulas, which can leverage over 325 purpose-built functions

• Native sequencing of dependent formulas

• Mixed-dimensional and mixed-grain operations

• Contextually-aware lookups to any level of detail, whether stored or calcu-lated

• Aggregation rules (e.g. sum, avg, max, min, count)

• Layered filtering

• Procedure modeling

• Predictive analytics

• Clustering

BOARDBEAMoffersagroundbreakingsolutionto incorporateall thepowerofpredictiveanalytics into daily business operations and the decision-making process, making it easier than ever for anyone to quickly gain powerful business insights and take action.

Turn your history into business predictionExploitallthepowerofBOARDBEAM’sautomatedpredictivemodelingtoexecuteextremelyaccurateforecastsinafractionofthetimeandeffortcomparedtotraditionalanalyticsolutions.

Thanks to its automated modeling approach, BOARD BEAM can automatically evaluate the characteristics of each time series, producing a suitable model, and can run forecasts thousands of times faster than competing products.

Theresultscanberefinedbyaddingfurtherinformationtothescenariotoreflecttheimpactofexternalvariables(covariates)ontheforecast.

Run instantaneous cluster and segmentation analysisBOARDBEAMcanautomaticallydefineclustersofsimilardataobjects,byusingthek-meansmethodology.

Userschoosethenumberofgroupsandtheobservationsets(e.g.profitabilityandturnover)and BOARD creates k-means clusters and makes them available as dimensions for analysis in the standard BI and performance management environment.

The seamless integration between the clustering capability and the BOARD environment not only allows for the immediate creation of reports, graphs, and self-service analyses based on clusters, but also enables the use of clusters for forecasting or planning.

Harness the power of BOARD statistical functionsBOARDBEAMoffersasetofout-of-the-boxstatisticalfunctions,fromtraditionaldescriptivestatistics likemin/max,average, standarddeviation, toalgorithmsspecificallydesigned forbusiness analysis such as frequency, recency, dormancy and nascency.

With this seamless integration into the BOARD platform, users can instantly leverage these functions for building any type of analysis, dashboard or report.

Page 14: BOARD Technology Overview

26 27BOARD Engine BOARD Engine

3.3

3.3.1 The toolkit: the fastest way to build analytic applications

Applications (“Capsules” in BOARD terminology) represent the environment where end-users can view, interact with and analyze data.BOARD provides all the tools necessary to build, deploy and maintain these analytic applications. These tools work together within an orchestrated system to provide an intuitive andvisualexperiencefortheuser.BOARD’sexceptionalease-of-useallowsanorganization’snon-technical people to become developers, quickly building applications without coding, leveragingallthecapabilitiesofferedbythedatabasesandthebusinesslogic.

Capsules consist of a collection of screens, with each screen being comprised of one or moreBOARDobjects.Objectsincludefunctionalitylike:reports,graphs,maps,folders,filters,business rules, gauges and navigation menus. These objects can be freely positioned on the screenthroughdragginganddroppingandtheninteractivelyconfiguredbytheuser.

Applications

Drag and Drop123

Configure

Run

3.3.2 Storytelling

Users can create personalized versions of application screens (subject to security) and save them in the form of a presentation. Once created, the new presentation can be saved into a personal workspace and shared with individuals and workgroups who can then collaborate on the contents.

Presentations use “live data”. When presentations screens are opened, the data is re-queried and the presentations maintain the same interactivity that they have in the application environment, including drills, pagers, selections, etc.BOARD’spresentationfunctionalityoffersagovernedbutagileandcollaborativeenvironmentthat moves reporting from a pre-canned, rigid and top down discipline into an adaptive and collaborative user-driven process.

• Eliminates the need for high levels of technical expertise, giving deci-sion-makers the power to build their own applications

• Enables applications to be perfectly customized to meet the business needs

• Provides flexibility to respond to ever changing demands on the business

• Perfectly fits agile development methodologies

• Substantially reduces the complexity and costs associated with building, customizing and maintaining a comprehensive BI and performance man-agement environment

The objects are automatically synchronized with each other and with the database, allowing users to immediately query, visualize and interact with the application as they build it. The applications leverage BOARD’s powerful business rules engine where users can model any typeofbusinessprocessorcalculation(workflow,multidimensionaldataentry,allocations,consolidations, etc.) using functional business terms and logic.

Thebenefitsrealizedthroughthisapproachinclude:

Page 15: BOARD Technology Overview

28 29BOARD Engine BOARD Engine

3.3.3 In-context Collaboration

User collaboration is greatly facilitated through BOARD’s native chat functionality. Users can work together on a shared analysis while interacting in real time in a chat session. Theability todefineworkgroupscompletesthepicture:virtual teamscan instantlydiscusstheir findings, explore new insights and make quick collaborative decisions based on avalidated, up-to-date and holistic view of enterprise data.

Analysis SharingLive ChatConcurrent Report Authoring

3.3.4 Cognitive technologies

“Cognitive technologies” are more than just fancy words, they represent a collection of technologies that enable people to interact with computers like people interact with each other.Withthese,comeexcitingimprovementsinease-of-useanduserengagement.

BOARDoffers the followingcapabilities,which leveragecognitive technologies, tobringatotally new way to interact with data across the entire BOARD platform. These capabilities are available to all users to the entire platform, cutting across all databases and applications.

Natural Language Recognition (NLR) and Generation (NLG) To simplify users’ access to information, traditional drag-and-drop report creation has been augmented with the ability to directly talk with the system or to interact through search strings, like performing a web search. Behind the scenes, a powerful technology translates speech into words, words into queries, queries into meaningful reports and reports into smart descriptions of the analytical highlights.

The Cognitive Space assistant

Voice Recognition

Page 16: BOARD Technology Overview

30 31BOARD Engine Clients

Search Generated Analytics BOARD’s Cognitive Search feature makes the creation of reports and analysis as easy as using a web search engine. Users can simply type a question and BOARD, through a fuzzy search mechanism, translates this request into a machine-generated report. BOARD will also evaluateeveryexistingscreeninthesystemandprovidearankedandscoredlistofscreensbased how well they answer the question.

Narration BOARDcangeneratesmartdescriptionsofyourreports,deliveredinbothtextandvoice,that highlight the most important and interesting elements. Instead of a business user having to manually review data to spot key highlights and then write commentary, BOARD can automate the entire process, generating smart and engaging content that highlights the key facts represented in the report.

=

+

Build once, deploy anywhere

BOARDensuresaconsistentexperienceacrossweb,mobileanddesktopenvironmentssodevelopers can build a single application and deploy it to users on variety of devices. BOARD applications can be accessed through the following clients:

CLIENTS4.

• BOARD Web

• BOARD Mobile

• BOARD Office Add-Ins

• BOARD Windows Client

• Build once without coding, deploy everywhere

• Consistent user-experiences with laptops, tablets and smartphones

• Modern user interface supporting gestures

• Enhanced experience - HTML 5 technology provides more features to web users and supports interactive procedures

• No need for plug-ins

BOARD Web

BOARD Web leverages HTML 5 technology and was designed for high performance and gesture based interactions. The powerful combination of HTML 5 with the “toolkit” allows users to build applications, without any coding, for immediate on from any device. BOARDwillalwaysprovideacompellingandconsistentuserexperience,whetheraccessedfrom a desktop, a tablet or a smartphone.

Key facts

4.1

Search

Machine-Generated Analysis

Research through existing reports

Page 17: BOARD Technology Overview

32 33Clients Clients

BOARD Web is a zero-footprint, rich-internet client that runs on all popular web browsers, including:MicrosoftEdge,MicrosoftInternetExplorer,MozillaFirefox,AppleSafari,GoogleChrome,MicrosoftWindowsandAppleMacOSX.

BOARD Web allows users to:

Any application developed on the BOARD Web Client can be accessed through either the web or the BOARD Windows Client, making it possible to manage hybrid (web and Windows) deployments without having to build and maintain two separate environments.BOARDWebsupportsmash-upandcanmakeBOARDinputsavailabletoexternalapplicationsoracquireinputsfromexternalapplicationswithouttheneedtobuildintegrationadaptors.This feature, combined with the Single Sign-On functionality, enables BOARD to be integrated with any company portal.

BOARD Mobile

BOARD Mobile is a native environment for Windows 8 and iOS based tablets and was designed for high performance and gesture based interactions.

BOARD mobile allows users to:

• Access BOARD applications

• Edit reports and run interactive analyses

• Execute data entry

• Build applications (coming in Board 10.2)

• System administration and management (coming in Board 10.2)

• Access BOARD applications

• Run interactive analysis

• Perform data entry

4.2

Users canalsoaccessdata inofflinemodemaking itpossible toworkwithoutanetworkconnection.BOARD Mobile applications are built using the “toolkit” in the same way as standard BOARD capsules and allow organizations to deploy almost any type of application screen to a mobile device in a matter of minutes.

Consistent user experienceBOARDensuresaconsistentexperienceacrossweb,mobileandWindowsclientssouserscan enjoy the same familiar environment on various devices.

Mobile data entryBOARD Mobile supports online data-entry. Users can input data on their tablet, save and immediately have the changes reflected across all devices. This allows organizations toimplement mobile business processes for simulation, planning and forecasting.

Shared insightsBOARD Mobile makes it possible to share reports and visualizations between BOARD Mobile users by sending links. The link recipient will be able to access and modify shared reports and analyses, subject to their security rights. Furthermore, any report or visualization can be easily shared as a static report on social networks through the native iPad/W8 functionalities.

Offline access to key dataBOARDMobileallowsuserstoworkbothonlineandoffline-userscanselectwhichreportsanddashboardstheywishtomakeavailableforofflineuse.

Rapid development and deploymentUsing the programming-free BOARD “toolkit”, mobile applications can be developed and deployed without a single line of code. All existingBOARDapplications created for adesktopcanbeeasily converted tomobileapplications. BOARD includes an automatic compatibility check to reduce development errors, time to solution and deployment issues.

BOARD Office Add-Ins

BOARD empowers users with self-service analysis and reporting within the MS Officeenvironment.Through the BOARDMSOffice add-ins, users can access BOARDmultidimensional datadirectly within Excel, Word and PowerPoint and also leverage traditional OLAP analysisfunctions,suchasdrill-down,sliceanddice,filteringandadhocquery.

BOARDobjects(e.g.reports,chartsandgauges)canbeeasilyembeddedwithinMSOfficedocuments, making it easy to create auto-updating booklets, factsheets and presentations.

ThedataflowbetweenBOARDdatabasesandOfficedocumentsisstreamlinedbyanautoupdate functionality that allows users to schedule data refresh at their convenience, ensuring full automation of the entire process and elimination of any manual activity.

TheBOARDExceladd-inalsosupportsdataentry inonlineandofflinemode,offeringaneffectiveoptionformanagingdatacollectionprocesses.Users can work without being connected to BOARD and automatically synchronize data whentheyfirstreconnect.Allofthisisdeliveredwithinthesecurityanddataconsistencyofthe BOARD platform.

4.3

Page 18: BOARD Technology Overview

34 35Clients Security

BOARD Windows Client

ConceivedforofferingarichWindowsexperienceandallthepowerofathick-client,BOARDWindows Client is the most comprehensive BOARD user interface.

BOARD Desktop allows users to:

A modern auto update functionality makes it easy to manage BOARD Desktop clients in large scale deployments, eliminating all the technical issues usually associated with Windows client versioning.

• Access BOARD applications

• Run interactive analyses

• Perform data entry

• Build applications

• System administration and management

4.4

APPLICATION SECURITY• Capsules Access• Capsules Editing

DATABASE SECURITY• Cubes (Measures)• Members• Single Cell

• Consumer/Builder/Admin• Read Only/Data Entry• Export/Print/Mobile

ROLE-BASED PROFILE

USER’S AUTHENTICATION• BOARD• Windows Authentication• LDAP• SAML (PKI) 1

23

4

Authentication

Authorization

SECURITY5.BOARD’s multi-layer security model is built to meet the most critical security requirements of any organization. With BOARD, you can combine data access authorization with application privileges and action permissions to provide pervasive, granular, cell-level security that can be easily personalized for individuals, user roles, and user groups.

Multi-tier security model

BOARD empowers users with self-service analysis and reporting within the MS Officeenvironment.Through the BOARDMSOffice add-ins, users can access BOARDmultidimensional datadirectly within Excel, Word and PowerPoint and also leverage traditional OLAP analysisfunctions,suchasdrill-down,sliceanddice,filteringandadhocquery.

5.1

Page 19: BOARD Technology Overview

36 37Security Security

Authentication methods

The authentication layer supports multiple authentication methods which can be selectively enabled or disabled:

• BOARD authentication: Thisisthenativebuilt-inauthentication.AnadministratorcancreatenewBOARDaccounts(i.e.username/password)anddefinerulessuchaspasswordexpiration,minimumlengthandmaximumretries.

• LDAP authentication: TheusernameandpasswordgivenbytheuserisvalidatedagainstanLDAPserver.ThismethodonlyrequiresconfiguringaconnectionbetweentheBOARDserverandtheLDAPserver,forexampleLDAP://myserver.mydomain.com:389/un=username,dc=mydomain,dc=com.Thismethodalsosupportscross-domainauthentication.

• Windows Authentication: InaMicrosoftWindowsnetwork,theBOARDservercanusethedomaincontrollertovalidatetheuseridentity.Whenthismethodisused,theBOARDClient,Officeadd-inclientorwebclientcanreusethecurrentWindowssessionauthenticationprovidingasinglesign-onwiththeactiveWindowsuser.

• PKI authentication (SAML): Withthisauthenticationmethod,theBOARDClientusesSAMLprotocoltoauthenticatetheuseronanIdentityProviderserverandpartofaPublicKeyInfrastructure(PKI).APKIauthenticationtypicallyprovidessmart-cardsorone-timepassword(OTP)authenticationmethods.

5.2

BOARD Authentication methods

Application authorization

The BOARD Capsules can also embed additional authorizations (e.g. to allow opening or editing a Capsule). This level of authorization ensures that a Capsule can’t be opened or alteredevenifitiscopiedormovedtoadifferentenvironment(i.e.adifferentBOARDserver)where the user may have a higher set of privileges.Theapplicationlevelauthorizationscanalsobeusedtorestrictexecutionofcriticalactionssuch as running a certain procedure that processes data in the solution.

Data access

TheauthorizationsdefinedonthedatamodelarestoredintheBOARDDatabaseandallow:

5.4

5.5

• Granting/denying access to create or modify a data model: authoring dimensions, cubes, data sources and other database objects

• Granting/denying access to cubes with read-only or read-write privileges

• Setting filters on dimensions down to cell-level (e.g. to limit a certain se-curity profile to a subset of data using the entities of the data model, such as regions, companies, cost centers or divisions

Role based authorization

AsetofauthorizationsisstoredasaBOARDsecurityprofilewhichcanbeassignedtoindividualuseraccounts.Followingtherole-basedsecurityparadigm,definitionsofauthorizationsarenotdefinedatindividualuserlevelbutasarole.There are two main sets of authorizations which form a security profile: one definesauthorizations on general features of the environment such as the Power User or Developer featuresandtheotherdefinesauthorizationsondata,suchascubesandfiltersondimensions. InaBOARDsecurityprofileitispossibletodefineauthorizationto:

• Database and Capsule authoring features through the license profile

• Restrict or grant access to given Capsule

• Restrict or grant access to a BOARD database, specifying a profile for each database

• Restrict or grant access to filtering, printing and other

5.3

Page 20: BOARD Technology Overview

38 39Security Infrastructure

Advanced security management

BOARD offers the possibility to store and manage authorizations within a BOARD“authorization” database that contains the metadata of all other BOARD databases. A simple idea - use BOARD to manage information (metadata) about BOARD databases and applications and build a completely bespoke solution to handle security using BOARD tools and features.Thiscapabilitynotonlymakesusermanagementeasyandefficient,italsoallowsthedelegationof security administration to key users within a BOARD application, making it possible to decentralizethesecuritymodelandtoperfectlyadaptittocomplexorganizationalstructures.

This enterprise feature uses a special BOARD database where cubes are used to grant authorizationsandaccessrightsoftheSecurityProfilesorofUsers.ThispermitsSegregationof Duties (SoD) where more than one person is required to authorize an account or a given authorization level.Creation, deletion, modification of authorizations or authorization levels are logged forauditing purposes.

5.6

BOARD Advanced security management

Cloud or on-premises

On-premises, on-hosting or on-cloud: you can easily implement BOARD with the same reliabilityandsecurity.WithBOARD,youcanchoosethemodelthatbestfitsyourbusinessatthepointofroll-out,havingthefreedomtomovetoadifferentmodelifandwhenyourneeds change.

BOARD Cloud is a SaaS version of the BOARD all-in-one platform. Backed by Microsoft Azure, BOARD Cloud gets your BI and CPM applications up and running faster while providing world-class security, reliability, scalability and performance.

Designedtoofferaseamlessconnectiontoon-premisesandclouddatasystemsinrealtime,BOARD Cloud allows you to deliver enterprise analytics, planning and forecasting that fully leveragesyourexistingsoftwareinvestments.

Unlike other solutions on the market, BOARD also supports writing-back from cloud to on-premises, through a secure tunneling connection. This way, your planning data stored in BOARD can easily be integrated back to the underlying legacy systems residing behind your firewall.

6.1INFRASTRUCTURE6.

RELATIONAL DB, DATA WAREHOUSE

BIG DATA, NoSQL DATABASESOn-Premise

data sources

Clouddata sources

...

ERP, CRMLEGACY SYSTEM

OLAP CUBES

EXCEL, CSV, TXT

BOARDCloud

Page 21: BOARD Technology Overview

40 41Infrastructure Infrastructure

In-memory cluster: horizontal scalability

BOARDoffersanadvancedin-memoryserverclusterarchitecturemakingitpossibletoshareBOARD server workload on a virtually unlimited number of nodes. This functionality provides horizontal scalability, supporting both and read and write operations.BOARD Cluster provides cutting-edge solutions to four main architectural issues usually experiencedinlargeandgeographicallydistributedimplementations,namely:

User scalabilityWhenthenumberofusersbecomessolargethatitaffectstheperformanceofasingleserver,BOARDClusteroffers thecapability toautomaticallyallocateusers to theavailablenodesthrough a load balancing mechanism.

Data partitioningThe management of big data projects often requires handling and analyzing very large data volumesthataffectperformanceofasingleserver.Inthiscase,BOARDClustercanbeutilizedtopartitiondataondifferentnodeswhilemaintainingaconsolidated“SinglePointofData”.Through an automatic sharing process, the BOARD database is transparently partitioned across nodes, allowing it to scale-out read-write BI and performance management applications without requiring any changes to them.

Process scalability and 24/7 availabilityThe deployment of large worldwide solutions with users working around the clock and in different time-zonesrequires thesolutiontobeavailableatall times, leavingno idle timewindow for performing routine overnight operations such as data feeds and bulk loads from source systems.

Key facts

6.2

• Sign in through a unified portal: cloud.board.com

• Private BOARD server on Microsoft Azure data centers

• Modern, enterprise-grade cloud infrastructure, security and reliability; including physical isolation of your data

• Hybrid architecture. Whether your source data resides in the cloud on on-premises behind your firewall, BOARD Cloud allows you to easily access and integrate them.

• Secure connection to on-premises data. The BOARD Cloud Connector does not require any firewall changes and immediately creates a secure tunnel between the BOARD Cloud server and your on-premises data.

• Write back to relational sources

• Native data connectors for the most popular cloud data sources

Inthiscase,theBOARDClusterallowsthesystemarchitecttodistributetheexecutionofbulkloadsorothermassprocessesacrossdifferentnodesofthecluster.Thisnotonlyoptimizesthe overall system performance by scaling out processes on different servers but it alsoisolates users from the impact of running mass data processes.

Geographic latencyPerformanceofworldwideorlong-distanceprojectsareoftenaffectedbyaproblemknownas“GeographicLatency”or,moresimply,infrastructuraldeficienciesinthenetworkconnection.

BOARD Cluster helps companies minimize these kinds of problems, making it possible to handle local data, users and operations on local nodes while having all the outcomes consolidated in a “Single Point of Data” through an immediate in-memory update.

This way, BOARD can ensure fast read and write performance to decentralized users, without impacting the application design or the overall functioning of the system.

SLAVENODESCLIENTS MASTER

NODE

BOARD In-memory cluster: architectural schema

Page 22: BOARD Technology Overview

42 43Infrastructure Infrastructure

• Users: BOARDMasterServerautomaticallyredirectsuserstothevariousslavenodes.Thisapproachissuitedtohandlealargenumberofuserswhentherearenootherspecificarchitecturalrequirements.

• Security Profiles: BOARDMasterServer redirectsusers toaparticular slavenodebasedontheirsecurityprofiles.Thisapproachisutilizedtomanagediffer-entkindsofoperationsondifferentnodes(i.e.delegatetoasingleserverallthedata-readingprocesses)ortohandlegeo-clustering(i.e.addressingalltheusersconnectingfromaspecificzonetoalocalserver).

• Performance Index: BOARDMasterServerredirectsusersonaparticularslavenodebasedonaPerformanceIndex,definedbythesystemadministrator.Thisapproachisutilizedtoallowthesystemadministratortofine-tuneusers’work-loadsbasedontheperformanceoftheunderlyinghardwarearchitecture.

• Ad-hoc configuration (Specialized Nodes): Whenmanaginggeographicallydistributed,complexplanningandsimulationprocesses,anautomaticload-bal-ancingconfigurationisoftenanineffectivearchitecturalchoice.Thesystemad-ministratorneedstheabilitytodefinewhichdata,usersandprocessestohandleonanysinglenode inorder toalign thesystemarchitecturewith thespecificfunctional requirements, resolve potential workload bottlenecks and optimizetheperformanceoftheentireprocess.

6.2.1 How the BOARD Cluster works

The BOARD Cluster is based on a hub-spoke model where a master node serves as a single point of access for a dynamic number of slave nodes.The master node addresses users, processes and data to slaves based on the policy implemented by the system administrator. The slave nodes work locally on their in-memory databases and send delta updates of data to the master node. The master node collects and merges data received sequentially (last one wins) and, after merging, pushes updated data to all slaves.

Atagivenpointintime,thedataonaslavenodecanbedifferent(morerecent)thanthedata in another slave node but eventually all the slave nodes become consistent (Eventually Consistent Model). The whole process is performed in RAM using the BOARD HBMP in-memory technology, thusmaking itextremelyefficient:eachsinglenodecan takechargeof its own data fetch and calculation processes (the most time-consuming activities) while only the delta data outcomes are synchronized between servers through a rapid in-memory update.

6.2.2 Adaptive load balancing

BOARDClustercanbeconfiguredinseveralwaystoadapttodifferentusecases:

BOARDWeb Mash-up

Interoperability and portal integration

6.3.1 Web mash-up and portal integration

The BOARD Web Client is a browser-based environment which can run BOARD applications directly on the web.

ThroughJavaScript it ispossible tocall themethodsexposedby theBOARDWebClientand interact with BOARD applications. These capabilities combined with the single sign-on functionality enables BOARD to be integrated with any company portal.

6.3.2 SharePoint integration

BOARD Web and SharePoint are both based on a Service Oriented Architecture (SOA) and Microsoft .Net Framework. This technology alignment makes the integration between the two systems quick and easy.

6.3

Page 23: BOARD Technology Overview

44 Infrastructure

Theintegrationcanbehandledatthreedifferentlevels:

Web page viewerThe simplest option to render a BOARD web screen on a Web Part Page is to use the Page Viewer. The Page Viewer is a default Web Part in SharePoint server and SharePoint foundation.TheSharePointadminsimplyentersahyperlink,filepathorfoldernametolinktothe content. The linked content of the Page Viewer Web Part is isolated from other contents on the Web Part Page by using an HTML IFRAME element.

This approach ensures that any HTML elements that are displayed as content in the Page ViewerWebPartdonotconflictwithotherHTMLelementsontheWebPartPage.The Page Viewer Web Part displays content asynchronously from the rest of the page, disconnecting the page performance to the linked object performance. This means that users can view and utilize other Web Parts on the page, even if the link happens to take longer to return the content than the other parts.This solution allows users to interact directly with the BOARD screen, but it is not possible to linktheBOARDenvironmenttoanySharePointobject,nortodefineanyaccessparameter(BOARD application is consumed “as is”).

HTML page integrationSharePoint allows users to visualize and directly access html pages.Through a traditional HTML Editor or Microsoft SharePoint Designer, it is possible to develop a web page embedding the html code needed to call the BOARD Web Client. This mashup solution allows the dynamic interaction between the BOARD Web Client and the web page.

ThehtmlpagecanbeacreatedinaSharePointlibraryorinanexternalwebsiteandthenretrieved from that point.This solution permits a comprehensive integration between BOARD and the SharePoint environment but will require the use of HTML and JavaScript programming.

6.3.3 BOARD web services

BOARDWebServicesallowflexibleintegrationofBOARDwithexternalapplications.BOARD was designed to be used as a web service and manages SOAP messages (usually formattedinXML)andreturnsXMLdatastructures.

The services supplied by BOARD can be used by any application in a simple and standardized way (W3C Standard), including: other web services, legacy systems, Java, .NET applications, web applications and by company portals.

Byprovidingtheseservices,BOARDcanextendadvancedBIandperformancemanagementfunctionstotheexistingcompanyapplicationenvironment.BOARDWebServicesarebasedon two types of methods:

• Data dictionary functions: (database metadata browser) that provide the list of Cubes (Measures), Dimensions, Hierarchies and Members

• Query execution functions: allowing access to data

BOARD has a worldwide reseller and partner network.For your local BOARD contact please refer to: BOARD International Partners

[email protected]

EUROPE / APACPiazza Bernasconi 5, 6830 Chiasso, SwitzerlandTel: +41 916975410Fax: +41 916975429e-mail: [email protected]

BOARD INTERNATIONAL Headquarters

ASIABOARD Asia Pacifi c60 Paya Lebar Road Paya Lebar Square, #08-23 Singapore 409051Tel: +65 62789138Fax: +65 63851690 e-mail: asiapacifi [email protected]

AUSTRALIABOARD Australia100 Walker StreetNSW 2060 North Sydney Tel: +61 289049777e-mail: [email protected]

DACHBOARD DeutschlandSchaberweg 2861348 Bad Homburg v.d.H.Tel: +49 6172171170Fax: +49 61721711770e-mail: [email protected]

FRANCEBOARD France21 Rue Auber75009 ParisTel: +33 143873736e-mail: [email protected]

JAPANBOARD Software K.K.Level 27 Shiroyama Trust Tower 4-3-1 Toranomon Minato-kuTokyo 105-6027Tel +81 03-5403-4660Fax +81 03-5403-4661e-mail: [email protected]

MEXICOBOARD MexicoDurango 269, Piso 3. 06700 Col. Roma NorteTel: +52 5585968693Fax: +52 5552720178e-mail: [email protected]

HONG KONG - CHINABOARD Hong KongUnit A, 20/F9 Queen’s Road Central, Central香港皇后大道中9號20樓A室Tel: +852 31897087Fax: +852 31897689e-mail: [email protected]

MIDDLE EASTBOARD Middle EastLevel 29, Marina PlazaDubai Marina,Dubai,UAEP.O Box 112229Tel: +971 4 559 5651e-mail: [email protected]

INDIABOARD India602, Sixth Floor, Landmark PllatinumPlot No.337/338Khar West, Mumbai 400052 Tel: +91 22 26050100e-mail: [email protected]

ITALYBOARD ItaliaVia Caduti di Marcinelle 520134 MilanTel: +39 022108071Fax: +39 0221098550e-mail: [email protected]

SPAIN - PORTUGALBOARD IbéricaAvda. Portal de l’Àngel, 3608002 BarcelonaTel: +34 93 4925702e-mail: [email protected]

UK - IRELANDBOARD UKTower 42, Floor 2025 Old Broad StreetLondon, EC2N 1HQTel: +44 2037272222 Fax: +44 1908255891e-mail: [email protected]

USA - CANADABOARD Americas33 Broad Street, Suite 200Boston, MA 02109Tel: +1 6172032173 Fax: +1 7812070347e-mail: [email protected]

AMERICA33 Broad Street, Suite 200Boston, MA 02109Tel: +1 6172032173 Fax: +1 7812070347e-mail: [email protected]

Where to fi nd BOARD

ARGENTINABOARD ArgentinaMaturin 2870Ciudad de Buenos Aires (C1416CTP) - ArgentinaTel: +5491164193822e-mail: [email protected]

BENELUXBOARD BeneluxKarspeldreef 141101 CK AmsterdamThe NetherlandsTel.: +31 20 7630591e-mail: [email protected]

Page 24: BOARD Technology Overview

www.board.com

Amsterdam,Austin,Barcelona,Boston,BuenosAires,Chiasso,Dubai,Frankfurt,Lima,London,Madrid,MexicoCity,Milan,Mumbai,

Munich,Paris,Singapore,Sydney,Tokyo

EN17

11-U

REQUEST A DEMO!