1 / 61

Taking Oracle Essbase to the Cloud

Taking Oracle Essbase to the Cloud. Kumar Ramaiyer Vice President OAC – Essbase - Product Development October 01, 2017. Essbase in Cloud (OAC): Design Principles. Architecture. Easy move to Cloud. Excel to Cube in secs. Thin client editor. Best MOLAP engine. Design in Excel.

stoned
Télécharger la présentation

Taking Oracle Essbase to the Cloud

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Taking Oracle Essbase to the Cloud Kumar RamaiyerVice PresidentOAC – Essbase - Product DevelopmentOctober 01, 2017 Confidential – Oracle Internal/Restricted/Highly Restricted

  2. Confidential – Oracle Internal/Restricted/Highly Restricted

  3. Essbase in Cloud (OAC): Design Principles Architecture Easy move to Cloud Excel to Cube in secs Thin client editor Best MOLAP engine Design in Excel Integration Usability Stability and Performance API and Headless Oracle Open World Session

  4. Architecture

  5. Proven Value, Redesigned for the Cloud • Re-architecture of Essbase for the Cloud • Designed to support higher concurrency and faster recoverability • Eliminated security bottlenecks • Memory utilization • Dynamic Scalability • Unified middleware layer for all services • Based on industry standards for ease ofintegration • J2EE Platform allows for quick innovations APS Scenario Management DBX Jobs EssbaseServer Thin Client Editor Dynamic Filters Data Source Grid UI Java Agent Catalog Federated Drill Through Scripting: R, Groovy, MaxL Java API, XMLA and REST Services Oracle Open World Session

  6. Architecture: Re-designed for Cloud on Industry Standards Cloud Portal Web UI REST/Java API REST Command Line Tool DVD Smartview SSH Cloud Network HTTP/HTTPS HTTPD server DBaaS Command Line Tools (MaxL, EssCMD) Essbase J2EE Cloud Platform Essbase JAgent DVCS, BICS Cloud Storage Service Cube Cube Cube Block Storage

  7. PaaS Benefits • Platform as a service -- offered as metered and non-metered • Buy only the capacity you need or as much capacity as you need. You can scale up and down -- elasticity is available for memory, disk and CPU • DBaaS, Cloud Storage, Block Storage and Compute -- all configured to buy and pre-requisites • Fully provisioned machine • Installation, Configuration, Patching and Rollback of patch • Backup and Restore • Scale up/Scale down [Scale out is on the roadmap] • Create/Start/Stop service • Monitoring of the service

  8. PaaS Benefits Variable Workload by time Different number of peak users at different time Use REST API to scale up/down memory and CPU Example: You have 16 OCPU box. At 9PM you can scale down to 1 OCPU till 4AM. Confidential – Oracle Internal/Restricted/Highly Restricted

  9. Easy migration to cloud

  10. Migration: Easy to move to Cloud • We deeply care about our existing customers and strongly encourage them to consider moving to cloud • It is still your favorite Essbase and much more  On-Premise ++++ • Your current on premise cube and all its features will run in the cloud • How to migrate? • Use LCM Utility to export all the artifacts • Use REST command line tool to import all the artifacts • Use DBX export tool to convert outline to DBX Excel workbook and build cube in cloud • We support migrating from 11.1.2.3 and 11.1.2.4 • We will consider migration from older versions if there is request and we welcome ideas and support and innovations from the community

  11. Excel to Cube in Seconds

  12. Modeling: Excel Automatic Analysis and Transformation Analyze and Transform Assign Time and Time Balance Measure Hierarchies User Input Measures Infer Hierarchies Use Column Hints Infer Attributes Enrichment and Dim Library Drill through for agg and skipped Text Measures Generate Future Periods Adjust Config for caches, etc. Allocation and Forecasting Automatic Solve Order Load rules and load Dim and Data XREF and Join with Sheets Support Formulas

  13. Modeling: Joyful and Simple • Takes us back to our roots as a Spreadsheet Database • Excel to Cube -- analyze and transform table to Cube • As simple as creating Pivot Table • Our Analyze and Inference abilities are improving almost daily and we are learning continuously • Many transformations will be applied automatically • Generated Cubes will incorporate more functionality with each patch • Goal: Bring in-depth Essbase analytics to new Excel users without having to learn multi-dimension modeling

  14. Design Cubes in Excel

  15. Modeling: Design by Example (DBX) • Excel “Contract” or “structured Excel workbooks” that represent a cube • Normalized breakdown for Dimensions and other Essbase constructs • Offline Modeling • Use power of Excel to generate the workbook • Cube Designer to author the Excel “Contract” • Use as Documentation • Easy to copy, modify and distribute -- Application distribution • Export to DBX and Import from DBX • Goal: Use Cube Designer and DBX Contract to model an Essbase Cube and incrementally update the Cube

  16. Design by Example (DBX) Workbook: Authoring Cube Designer LCM Utility on On-Premise Cubes Excel Contract (DBX) Copy and Edit Workbooks Export from UI Gallery Templates and Dimension Library Export from Cube Designer

  17. Our Guiding Principle … • Essbase in the Cloud • Automatic creation of cubes from Excel • Centralized and Secure Server Based Modeling • Easy to enforce business drivers • Easy to feed from/into other systems • Easy to collaborate and support workflow • Scalable as the model grows • Efficient aggregations and calculations over massive data store: not just what is visible and not limited to 1 M • Easy to assemble plans from multiple apps and aggregate those plans – integrated business process • OLAP modeling with dimensions, and hierarchies • Accessible for other applications to add analysis and planning capabilities over additional transactional data • Scenarios and Sandboxing are supported for isolation • Support DV user interface as well as Excel interface • Customers Like Excel • Easy to use • Provides isolation • Expressive calculations • Full Dependency analysis • Automatic Data flow after a change • Easy to customize new changes to metadata With OAC, We allow them to keep all these benefits of Excel but also add Oracle Confidential – Highly Restricted

  18. Best MOLAP Engine – Maintain the edge …

  19. Hybrid: Dynamic Calculations and Aggregations • Biggest innovation in Essbase history • Brings power of BSO calculations and power of ASO dynamic aggregation and calculation to a single engine • Unified query processor for BSO and ASO • Essbase Cloud • Hybrid BSO is default for BSO • When Excel is imported, all upper level and formula members are set to dynamic • Customers can disable Hybrid at application level • Calculation scripts use Hybrid when accessing Dynamic members • Aggregation views for Hybrid BSO is in the works -- use intermediate stored levels as an alternative

  20. Dynamic Calculations and Aggregations -- Benefits • Size of database shrinks exponentially • Pre-computation and Over-computation is no longer required • All computations happen on-demand and in-memory • Lot less IO: only for loaded data and for user lock-and-send • Restructuring time reduces dramatically as little data to restructure • We generalize calculation order using ”solve order” • Helps with Excel to Cube -- easy to convert Excel formulas to dynamic formulas. Generation of calculation scripts is harder and not efficient • Calculation script -- reduced to procedural activities like data copy, data clear, data export, etc. • Backup Size and time, Disaster Recovery and High Availability -- all benefit

  21. Other Engine Improvements • Shipped with configuration (essbase.cfg) that is optimized for a given POD • Default functional parameters reflect our vision for future direction of essbase in the cloud • Hybrid, Idle-time shut down, Automatic solve order, All kinds of caches optimally tuned • Users can override these at Application level • In-place data write and lockless algorithms • No fragmentation of page files • cloud friendly -- no need for frequent restructuring • Member ID usage for non-unique outlines • Helps to keep smart view sheets even when members move to different hierarchies

  22. Usability

  23. Usability • Auto generation of cubes • Formulas made simple -- no runtime error for bad formulas • Error messages are very detailed and pointed • Query tracing -- provides detailed steps and order of calculations • Calculation tracing -- for any POV, you can run a calculation script and ask for explanation of how the engine computed that value • Data Audit Trail -- Gives detailed breakdown of who changed what POV at what time • Metadata Audit Trail -- Gives detailed breakdown of who changed what metadata at what time

  24. Usability • In Smartview, for any dynamic cell, you can check for details. It shows all the dynamic members for POV and corresponding solve order and what operation was used to compute the cell • In Smartview, you can change solve order of that cell -- first time metadata update • Job Queue -- Asynchronous execution and gives details and one can look at history of calculation, load, dimension build • All Logs are accessible from UI at Application level and at global level • Location Aliases for XREF and Partition details (host, user, password) for local sources are not needed

  25. Usability Goals • Help users to get “unstuck” and proceed further analyzing • Expose the engine behavior in a consumable form -- calculation tracing of scripts and solve order details of dynamic POV • Move from one valid state to another valid state

  26. Stability and Performance

  27. Stability and Performance of Essbase in the Cloud Enterprise Planning Cloud Enterprise Performance Reporting Cloud Planning & Budgeting Cloud Financial Consolidation & Close Cloud Sales Cloud Project Management Cloud Financials Cloud • Oracle Confidential – Internal

  28. Stability and Performance in the Cloud • Supporting SaaS Applications in the Oracle Cloud for close to 5 years • EPM Cloud uses Essbase extensively for 3+ years • In SaaS, no tuning or configuration changes are allowed … yet Essbase is highly stable and performant • Essbase in PaaS builds on top of that code base and adds • Relational store for Security as well as J2EE platform, REST, http protocols and other standards • Cube folder under app/db to cleanly and automatically recover from restructure issues • Hybrid and in-memory computation • Idle-time shutdown to free memory for other active cubes • Resource guards for memory, CPU and disk -- gets into maintenance mode • DR support with RPO of zero and RTO of 2 to 4 hours

  29. API – Headless Access

  30. API -- provide complete access • Two methods • Java API over http(s) • REST API • We consume these interfaces for all our tools • Smart view queries use Java API • Command line tool uses REST • Cube Designer uses Java API and REST • Our new JavaScript UI is completely based on REST

  31. New Command Line Tool to access Essbase in the Cloud • Execute jobs such as calculations, clear data, data load and dim builds • Upload and Download files to the cloud service • Check job statuses • LCM Import/Export • Execute MaxL jobs • Remote Data Push using http • Backend integration • Uses REST API Oracle Open World Session

  32. Integration

  33. DV and Essbase - Complete Offering • Essbase • Allow users to interact with the data using Excel • Ingests and harmonizes spreadsheets, collates and manage disparate inputs • Provide coherence, consistency and sharing • Enables and manages the iterative information cycle • DV • Visually research and discover • Review and visualize personal data • Review and visualize group data and in the context of wider corporate data • Provides insights at key stages of the iterative information cycle • Automatically creates an Essbase cube when needed • Excel, BI, Source Systems • Data from any source • Exports from other systems • Hand crafted spreadsheets • Essbase – Data Interaction and Collaboration • Interact with the data and model using Excel with a strong server side data consolidation and integrity (multi-user read/write access) • Add model validation and shared KPIs for user scenarios and data gathering • DV – Visualize and Discover • Visual discovery • Share analysis and results • Track and monitor progress • Create an Essbase model for multi-user write collaboration, scenario management and What-If analysis • Oracle Confidential – Internal

  34. Cloud Console -- Common PaaS Tasks • Create service • Interactive and easy process, different host types • Restart/Stop/Start Service • At any time user can choose to restart VM, stop it or start it again depend on the usage • Scale Up/Scale Down • If you need a bigger host, in one click you will get it • Backup/Restore • At any time you can trigger full data backup process or restore already existing backup • Patch/Rollback Patch • You can get patches in one click, you choose when and what • Monitoring • Get status of your service

  35. Where are we going next …

  36. Essbase in Cloud -- Roadmap Infrastructure • Scale-out of Essbase • TCP/IP replaced by http(s) • Scheduling at PSM and Jobs Core Server • Tuple Based Calculations • Dimension Properties • Multi-parent hierarchy with weights • Unstructured Export from Cubes • DR Support • Compression of data in API and CLI • Partition to REST API sources Platform and UI • Incremental Unstructured • GRID UI in Web • Scenario Management in SV • Validation of filters in Web Confidential – Oracle Internal/Restricted/Highly Restricted

  37. Take Home Message … • Essbase in the cloud -- thoughts for you to carry home • Simple and Easy to Use • Highly secure and highly scalable • Solid architecture and Lot of innovations at rapid pace • Help Excel Users to easily move to Cloud for • Analytics, • Collaboration, • What-If and Scenario Modeling • Simple and Strong Integration with DVCS/BICS and DVD • We maintain the edge as BEST MOLAP Engine in the world • Faster and faster query execution, calculations, load, restructure, …

  38. FAQ

  39. Client Centric Excel Universe to Essbase in Cloud Essbase in Cloud Deployment Modeling Integration Life cycle Confidential – Oracle Internal/Restricted/Highly Restricted

  40. Deployment -- Architecture Cloud Portal Web UI Rest API Rest Command Line Tool DVD Smartview SSH Cloud Network HTTP/HTTPS HTTPD server DBaaS Command Line Tools (MaxL, EssCMD) Essbase J2EE Cloud Platform Essbase JAgent DVCS, BICS Cloud Storage Service Cube Cube Cube Block Storage

  41. Deployment -- New Mid-tier • Re-architecture of Essbase for the Cloud • Designed to support higher concurrency and faster recoverability • Eliminated security bottlenecks • Memory utilization • Dynamic Scalability • Unified middleware layer for all services • Based on industry standards for ease ofintegration • J2EE Platform allows for quick innovations APS Scenario Management DBX Jobs EssbaseServer Thin Client Editor Dynamic Filters Data Source Grid UI Java Agent Catalog Federated Drill Through Scripting: R, Groovy, MaxL Java API, XMLA and REST Services Oracle Open World Session

  42. Deployment -- Repurpose Investment in Client Automation MAXL On Premise On Premise Network • Uses the C Agent • Uses the C API • Communicates using TCP/IP Server TCP/IP Essbase Runtime Client Essbase C Agent Client Cube MAXL Scripts Oracle OpenWorld Session

  43. Deployment -- MAXL using OPC VPN/Fastconnect OAC with VPN/Fastconnect • VPN allows OAC to become an extension of customer data center • Client RTC uses C-API • Communicates using TCP/IP • Need to open 1423 agent port and server ports for client and OAC communication • These ports need to be updated using OAC firewall policy controlled by the customer On-Premise OAC POD Essbase J2EE Cloud Platform Java Agent Client Essbase Runtime Client TCP/IP Cube MAXL Scripts Oracle OpenWorld Session

  44. Deployment -- MAXL using Compute Node OAC and Compute Node • Same as previous architecture • In this case, the client machine itself is moved to a compute node in the OPC • Like moving your client code/machine from one data center to another data center • Both compute node and OAC nodes are behind OPC firewall and are not exposed to public network Compute Node OAC POD Essbase J2EE Cloud Platform Java Agent Client Essbase Runtime Client TCP/IP Cube MAXL Scripts Oracle OpenWorld Session

  45. Deployment -- MAXL using EssCS • EssCS is a new REST based client tool built for Essbase cloud • It communicates to Essbase cloud using http(s) secure protocol that is industry standard and also supports TLS 1.2 • Customers need to re-write their scripts to use EssCS • But they can re-use their MaxL scripts On-Premise OAC POD Client Essbase J2EE Cloud Platform Java Agent Essbase CLI HTTP Cube MAXL Scripts MAXL Scripts Oracle OpenWorld Session

  46. Fast Connect Confidential – Oracle Internal/Restricted/Highly Restricted

  47. FastConnect : Overview • Dedicated:Access your Oracle Public cloud services in a secure, consistent and cost effective manner. • Reliable:Delivered as a fully redundant service with two physical connections from your network edge. • Standards Based: Leverages industry standard BGP routing to manage the exchange of routes between Oracle Public Cloud and your networks. • Rapid Service Provisioning: Service can be turned up rapidly (in minutes) if you are already in the same Datacenter. FastConnect Partner Edition

  48. FastConnect : Use Cases • Bidirectional transfer of large volumes of data (batch jobs) • Unpredictable Internet results in significantly lower performance or batch jobs not completing in time. • Adding Internet capacity to your corporate network adds up to the overall cost but with limited return or overall improvement in performance. • FastConnect overcomes this problem by migrating all OPC traffic to a dedicated path thus improving overall performance

  49. VPN for Compute: Overview • Simple: Within few steps, IPSec tunnel can be created from customer premise to Oracle Compute Cloud • Cost Effective: Eliminates the need to invest in leased line for more security • Secure: AES128 encryption secures your data from any network sniffers or attackers • Reliable: Redundant appliances that ensure your data is highly available Virtual Private Network Securely extend your Network to Cloud

More Related