More Related Content
Similar to Finit - What's Trending: Topics on the Cloud and Infrastructure (20)
Finit - What's Trending: Topics on the Cloud and Infrastructure
- 1. Key Topics on the Cloud and Infrastructure Impact to the CFO organization and EPM / CPM solutions May 13, 2016 Rob Cybulski
- 2. Why is Finit here? Page 2 Finit takes pride in being a company of makers and doers; people who dig in and get their hands dirty We make and deliver personalized financial reporting solutions and experiences …Not Company to Company, but Person to Person By listening, questioning, and analyzing
- 3. Finit Culture Slide 3 No debt or external ownership means complete alignment & focus on our client’s needs and vision We bring a quality of craftsmanship to our designs and execution that lead to solutions that are functional, usable and efficient We actively seek feedback and invest in training and professional growth in our consultants We use employees, not contractors. Compensation based on CLIENT SATISFACTION A CULTURE OF INTEGRITY ADVOCATES OF YOUR INTERESTS DRIVEN TO DELIVER TAILORS OF OUR CRAFT
- 4. Our Partnership with Clients Slide 4
- 5. Finit Customer Success Page 5 Our values, culture, and approach to becoming a trusted advisor to our customers has led to 100% customer success for every Finit client (250+) and for every EPM / CPM project (500+)
- 7. Questions
- 8. About the Presenter Slide 8 Rob Cybulski [email protected] • Co-Founder of Finit • 17 years of EPM / CPM Implementation experience • Assisted over 100 clients in various strategy, design, development and leadership activities.
- 9. Agenda • Cloud computing • Cloud computing service models • Cloud deployments • EPM / CPM vendors approaches to the Cloud • Deployment Use Cases • Infrastructure and Platform • Virtualization • Performance Drivers • Backup and Recovery Approaches • Technical Testing Strategies Slide 9
- 10. Cloud Computing
- 11. Cloud Computing Introduction • Cloud Computing: a network of remote servers accessible through the internet that are used to provide resources, data and / or applications when required (i.e. on- demand). • Cloud Computing can apply in many different forms and has different service models ranging from providing hardware / infrastructure resources to providing software applications. • Additionally, Cloud Computing can be deployed different ways, from having it publicly available to everyone to being privately available to just your company. • Prior to the Cloud, EPM / CPM customers used two methods: • On-Premises: where clients themselves procured, managed and hosted the infrastructure and software applications. • Hosted: where clients used a 3rd party that procured, managed and hosted the infrastructure and software applications. Slide 11
- 12. Cloud Service Models Slide 12 • Infrastructure As A Service (IaaS) • Pronounced as “EYE-AZZ” • IaaS provides the infrastructure and hardware resources (i.e. servers, storage and networking) through the Cloud. • Platform As A Service (PaaS) • Pronounced as “PAZ” or “PASS” • PaaS provides ‘platform’ resources such as databases and web services. It can also provide development resources when used for application development purposes. • PaaS and IaaS providers have some overlapping areas. • Software As A Service (SaaS) • Pronounced as “SASS” • SaaS provides software applications and solutions through the Cloud.
- 13. Infrastructure As A Service (IaaS) Slide 13 • At its core - provides hardware and infrastructure resources (i.e. servers, storage and networking) for companies to use, but the client configures the hardware with whatever software they want. Clients are essentially "renting" the hardware instead of owning it. • With an IaaS model, the client is typically responsible for all software management. • Client buys the infrastructure and typically has the ability to scale as needed: • Base setup may be 4 servers with 16 CPU’s, 32 GB of RAM and 500 GB of storage each. • May scale during peak periods to add more resources. • How it can be used in EPM / CPM: • Use IaaS to be the underlying infrastructure that your EPM / CPM software products are installed and deployed from. • Example of IaaS providers: Amazon Web Services (AWS), IBM, Rackspace
- 14. Platform As A Service (PaaS) Slide 14 • At its core – provides core foundation components such as operating systems, databases and web services. Collectively, the components are the ‘platform’ available to clients. • PaaS providers typically provide the platform along with hardware components available with IaaS providers. • Most PaaS providers provide some of the same hardware components in a traditional IaaS offering. • Outside of EPM / CPM, it can be used for software development by giving developers toolkits and a development execution environment. • How it can be used in EPM / CPM: • Use PaaS to be the underlying infrastructure and platform that your EPM / CPM software products are installed and deployed from. • Differs from IaaS as you get the databases, web services and foundation components installed and managed, whereas with an IaaS you typically need to set them up and manage them. • Example of PaaS providers: Microsoft Azure, Google App Engine
- 15. Software As A Service (SaaS) Slide 15 • At its core – provides software applications accessible through browsers and other third party tools (i.e. Microsoft Office). • With a SaaS model, the client pays for use of the application and the software vendor manages the infrastructure, platform, software and all tasks (upgrade, patching, backups, performance, etc.). • In most SaaS models, the client has no insight or knowledge of the infrastructure or platform, they just see the applications. • Subscription-based licensing is the norm in a SaaS model, and in many cases you pay on a per user, per application basis. • Multiple applications of the same product may involve multiple licensing subscriptions. • May not be best fit for Enterprise customers utilizing multiple solutions. • How it can be used in EPM / CPM: • Deploy solutions to users on SaaS solutions • Example of SaaS providers: Oracle Cloud, Salesforce, Adaptive Insights, Host Analytics
- 16. Service Model Summary for EPM / CPM Slide 16 • IaaS / PaaS • Model for deploying EPM / CPM software applications where a complete integrated server environment exists. • Software will continue to be traditionally licensed (named user +maintenance) • Client can deploy as many software applications / solutions as they want on their IaaS / PaaS cloud environment. • Client or 3rd party vendor required to install and manage the EPM / CPM software applications. • In PaaS model for EPM / CPM through a provider such as Microsoft Azure, client gets integrated platform items on top of IaaS infrastructure. • SaaS • Model for deploying pure Cloud EPM / CPM software applications. • Software licensed on subscription basis (per user per application) • Software vendor manages all infrastructure, platform and software components. • In many EPM / CPM SaaS models, solution fit is for more stand-alone requirements versus complete CFO solution.
- 17. Cloud Deployment Options Slide 17 • Public Cloud • A public cloud is available for public use. In a SaaS service model, your SaaS application will be specific to your organization, but will be deployed on shared infrastructure and platform resources. • Private Cloud • A private cloud exists only for one organization and can be hosted either internally or externally. An internally hosted private cloud is really just hosting your own servers. • An external private cloud is dedicated to you, but your data sits outside your firewalls. • Hybrid Cloud • A Hybrid cloud deployment involves two or more cloud deployments that remain separate. • In CPM / EPM terms, this may mean using SaaS applications via a public cloud for some solutions and applications via a private cloud / on-premises deployment for others.
- 18. Cloud Benefits for CPM / EPM Slide 18 • New Functionality Sooner -> More Frequent Upgrades • Software vendors managing SaaS & PaaS EPM / CPM solutions tend to upgrade solutions more quickly than clients on their own. • Quicker Deployments • Cloud solutions can be ready for use in minutes or days versus previous lengthy process. • Potential for Cost Savings • Cloud deployments and licensing models may overall lead to cost savings for organizations. • This may additionally be achieved through the scalability / elasticity of resources by not paying for peak resources all the time and paying for them only when you need them. • Scale for Performance • IaaS and PaaS with software designed for those models have the ability to scale and deliver tremendous performance on-demand.
- 19. Cloud Challenges Slide 19 • Less control / Privacy • Less ability / insight into how the solutions run and operate • More formal process to fix issues • Less ability to force change • All data is maintained on vendor data centers / outside your walls • Infancy of SaaS solutions • SaaS solutions are largely targeted for small / mid-market or stand-alone needs. • SaaS solutions can lack ability to scale for performance. • Integration is in infancy and in many cases, utilizes processes that existed 20 years ago. • Many vendors lack a comprehensive strategy for the Enterprise • Cost • In some cases, a subscription model is more costly over the long- term for an organization than a traditional licensed approach.
- 20. CPM / EPM Vendor Cloud Slide 20 Products/Software On-Premises IaaS PaaS SaaS OneStream Oracle Hyperion Legacy: HFM, Planning, Essbase Oracle Hyperion Cloud Solutions: PBCS, ARM, FCCS, EPR SAP Host Analytics, Anaplan, Adaptive Insights
- 21. Deployment Use Case - OneStream Slide 21 • IaaS or On-Premises / Private Cloud Deployment • OneStream deployed on servers either through an IaaS service provider or internally through on-premises environments. • Good use case for: • Clients who want full control of infrastructure and services. • Not good use case for: • Clients who want to lessen their IT responsibilities. • PaaS Deployment • OneStream deployed through Microsoft Azure and their PaaS service model. • Good use case for: • Clients who want to minimize IT responsibilities • Not good use case for: • Clients who want full control of infrastructure and their data.
- 22. Deployment Use Case - Oracle Slide 22 • SaaS Solutions Only • Utilize Oracle Cloud solutions such as PBCS, FCCS, ARM CS, etc. • Good use case for: • Small / mid-market organizations • Stand-alone or silo solutions not fully integrated with other CFO solutions • Not good use case for: • Enterprise clients who have multiple solutions across products • Customers who want more control on their environment and solutions
- 23. Deployment Use Case - Oracle • Hybrid Solution (SaaS + Other) • Utilize Oracle Cloud for some solutions (i.e. PBCS, ARM CS). Utilize IaaS or on-premises model for some other products (HFM, Essbase). • Good use case for: • Clients who require the functionality of HFM • Clients who have multiple Essbase cubes / applications • Clients looking to address a specific need with a solution (i.e. Account reconciliations) • Not good use case for: • Clients wanting a fully integrated EPM / CPM solution • Clients wanting full control of all of their solutions and data. Slide 23
- 24. Deployment Use Case - Oracle Slide 24 • On-Premises / IaaS Solutions • Utilize Oracle Hyperion products either through an on-premises deployment or an IaaS provider. • Good use case for: • Clients who require the functionality of HFM • Clients who want a complete integrated environment of all CFO solutions. • Clients with large numbers of applications • Not good use case for: • Clients that don’t want to manage Infrastructure or certain IT functions (backup, performance, etc.) • Small / Mid-Market organizations • Clients with a small number of application needs
- 26. Virtualization - Introduction • What is it and why are IT departments moving to it? • Virtualization means simulating computing resources on a physical host server (computers, storage, networking, etc.) • Significant benefits to IT • Faster server provisioning • Consolidation of resources • Spreading computing workload • Easier hardware upgrades • Disaster recovery and backup • Nearly all of our customers employing some level of virtualization • Most often used for non-production environments • Many production environments have virtual/physical mix • Some have corporate mandates to go 100% virtual Slide 26
- 27. Virtualization – Key Points Slide 27 • Oracle Hyperion and OneStream EPM / CPM products work on virtualized servers. The key consideration is stability and performance. • Stability challenges with virtualization stem from moving VM’s while EPM / CPM application processes are running. • Consolidation or large allocation calculations are very intensive activities generating significant volumes of data into the database. Moving that virtual machine across the network during those processes can cause performance issues or errors. • Performance challenges exist because: • VM’s add a layer of overhead to the base machine (7 – 20%) • VM’s may not have performance-grade CPU’s and RAM needed to maximize performance
- 28. Leading Practices for Virtualization Slide 28 • To maximize performance and stability in CPM / EPM applications using virtualized servers, we recommend the following: • Dedicated physical host server for the EPM / CPM servers • Avoid unscheduled vMotion or DRS events • Give physical host performance-grade CPU’s and RAM sufficient for complex analytical processing • Highly tune EPM / CPM application servers and RDBMS. • If the option exists, the biggest gain in physical servers would be in the application and RDBMS tiers.
- 29. Delivering Performance Slide 29 • Infrastructure • Across OneStream and Oracle Hyperion products, processors make a tremendous difference. • We have seen performance improvement of 30% strictly based on using different processors. The better the processors, the faster EPM / CPM calculations will run. • Virtualization / Physical • Virtualization adds some overhead (7 – 20%) • Storage / Network • Networking and speed of storage devices can greatly affect performance of intensive calculations. • Solution Design & Tuning • Solution design and tuning can greatly affect performance. • Varies by software product and vendor.
- 30. Backup and Recovery Strategies Slide 30 • Most EPM / CPM solutions for the CFO organization are considered very critical and require maximum up-time. • 2 separate strategies should be developed: • Backup: strategy and processes for taking backups or snapshots of the current CPM / EPM solutions and components, to be used to restore to the existing Production environment. • Disaster Recovery: strategy and processes for backing up and restoring CPM / EPM solutions and components to a different data center.
- 31. Disaster Recovery Strategies Slide 31 Our EPM / CPM clients have largely developed one of two Disaster Recovery strategies: • ‘Hot Standby’ Strategy • A separate live or available EPM / CPM environment exists in a separate data center. • Backups are shipped to this environment nightly. • In a disaster, backups are restored into the live environment. • Generally up and running within 4 hours or less • ‘Cold Standby’ Strategy • Backups of EPM / CPM virtual machines are shipped to a separate data center nightly. • In a disaster, VM’s are turned on in the separate data center. • Both environments cannot be live and running at the same time. • Generally up and running within 8 hours or less
- 33. What are Technical Tests? • They are tests that involve testing the infrastructure/platform/environment. • What are we trying to do in these tests? • Deliver Performance and a Great User Experience • Ensure users can access the solution(s) • Deliver the solution to their access points (web browser, Excel, etc) • Determine if the solution is running as fast as possible • Ensure Stability • Determine if the environment can continue to deliver a great experience as it reaches peak demand • Ensure the system remains stable through all types of user interaction and requests. • Be Prepared • Ensure we can recover if something goes wrong. Slide 33
- 34. Types of Technical Tests Slide 34 • Connectivity Testing – test to ensure connectivity, access and acceptable response times to the new server environments. • Performance (Stress / Load) Testing – test to ensure the solution can deliver stability and great performance and experiences at peak demand. • Recovery Testing – test to ensure we can recover from a backup (into existing environment) and to a new data center.
- 35. Connectivity Test Slide 35 • Test Objective: • Can users access the solution at their physical location using their access tools (IE, Excel) and get quick and acceptable responsiveness? • Test Scope • Perform and measure select application tasks. • Minimum of 1 user per physical location and can be an IT user • Risk Factors to Evaluate • The way users connect (Direct, VPN, Citrix) • Distribution of users (US only vs Global) • Similar systems being used or is this the first? • When do we conduct? • Generally early in the implementation to facilitate remediation time • Post Go-Live Testing • Generally not needed for existing locations, but should be incorporated for new physical locations
- 36. Performance (Stress / Load) Test Slide 36 • Test Objective: • Can we deliver a great experience as we reach peak demand and can the system hold up and be stable when activity is high? • Test Scope • Perform and measure select application tasks. • Usually system generated, not direct user involvement • Risk Factors to Evaluate • Expected Concurrency & User base • Size and complexity of applications • When do we conduct? • Post Install – Initial test to validate environment (Sample App) • End of Build – Second test to test real solution (Real App) • Post Go-Live Testing • Generally not needed, though may be incorporated when major application changes exist or for large deployments.
- 37. Recovery Test Slide 37 • Test objective: • Can we recover when something goes wrong? • Test Scope • Perform Recovery steps and then test application and validate data to ensure accuracy • Admin team / IT perform test • Risk Factors to Evaluate: • Number of products / applications • Environment stability • When do we conduct? • Prior to go-live: need real apps and full solutions to ensure process • Post Go-Live Testing • Should be fully tested once a year to ensure backups and procedures can be properly performed.
- 38. Thank You for Attending! Page 38 Today’s Presenter: Rob Cybulski [email protected] General Questions: Greg Barrett [email protected] Joe Mizerk [email protected] Copy of the slides or Recording: Email us for a copy of the slides or a link to the recording [email protected] Past webinars: www.finit.com/webinars Follow us on Twitter for updates: @Finit_Solutions