openUC & Virtualization
1 like1,021 views
This document summarizes a presentation about virtualization and how it relates to openUC and sipXecs. The presentation covered why organizations virtualize, criteria for running applications virtually, how recent versions of openUC/sipXecs are optimized for virtual environments, and case studies of virtual deployments. Recommendations were provided on virtual resource requirements for openUC installations of various sizes. A private cloud case study detailed a 500-user virtual setup across multiple virtual machines.
openUC & Virtualization
- 1. 1 Virtualization moderator: jerry stabile March 10, 2013 / Bentley Univesity / Boston MA
- 2. Topics 2 > Why Virtualization? > What makes a good Virtual App? > Cloud Computing Models > Operating Environment > Virtual Case Study
- 3. Intro of Presenters Team 3 Mike Picher Douglas Hubler Jerry Stabile
- 4. Why Virtualize? Can RTC take advantage of virtualization 4 > Virtualization is a key building block to cloud computing > Enabling technology that creates an intelligent abstraction layer which hides the complexity of underlying hardware or software > Drives the evolution of IT infrastructure to standardized compute platforms Infrastructure Consolidation Business Continuity Rapid Deployment Backup/Restoration
- 5. Criteria for running virtualized applications Software architecture considerations 5 >Realtime processing >Support flexible backup/restore >Distributed vs monolithic >Enhanced monitoring/right sizing >Easy to install
- 6. openUCTM release 4.6 readiness How R4.6 leverages virtualization 6 > Optimized media service for RTC > Redesigned backup/restore – CFEngine > Enabled more distributed services > Embedded monitor server > Automated installer > RPM based > Environment agnostic
- 7. Cloud Operating Environments Public, Private, Hybrid, Community Clouds? 7 Private Cloud Public Cloud IT Managed Maximum flexibility Secure, available Any device, any time, anywhere Heterogeneous Third party provided Less scalable Higher Risk – security, availability More expensive Monolithic Office workers P Confidential information Home workers Performance sensitive apps Mobile workers Hybrid clouds - partnership public and private cloud computing and services providers Cloud Services Cloud Computing
- 8. siopXecs/openUCTM Cloud Certification Program Communications as a service from the cloud 8 > Goals: > Become virtual environment agnostic > Leverage existing production environments > Utilize environment specific management tools > Enable enterprise application store > What is this program about? > Creating best practice implementation process and documentation > Providing technical assistance and training > Quality assurance through load testing automation and use case validation > Network infrastructure readiness consulting > Integration assistance for service provisioning and assurance
- 9. Operating environment requirements How to optimize a SW application for virtualization – what to look for? 9
- 10. openUC / sipXecs 4.6 What can we virtualize? 10 > openUC / sipXecs 4.4 and earlier > Could easily virtualize proxy/registrar servers. > Not media services > openUC / sipXecs 4.6 > Can virtualize entire system. > Why? > Red Hat / CentOS 6.x – Tickless Kernel > New Timer Modules for FreeSWITCH (our media services)
- 11. openUC Virtualization Planning 11 > Performance > Estimate 10 to 20% less performance in Virtual Environment > Host Servers > Minimize interaction between High CPU / High Bandwidth need virtual servers and openUC on same host. > Dedicate processor and RAM when able.
- 12. openUC Resource Footprint 12 > Minimal Configuration > 1 Core, 3.7 GB of RAM, 50 GB HD (AWS m1.medium) > Supports about 20 concurrent calls to media services (conf, vm) > Up to ~ 100 users. > Can make memory use less with some ‘tweaking’ of sipxconfig heap usage. > SIP Capture takes significant disk space, turn off for small installations. Bandwidth Utilization > Peak ~ 200 Kbps / 10 users (1 of every 5 users on phone)
- 13. Recommended Resources 13 > 0 – 75 Users – 1 Core, 4 GB of RAM, 80 – 100 GB storage > Up to 500 Users – 4 Cores, 8 GB of RAM, 200 – 300 GB storage > Up to 5000 Users – 8 Cores, 32 GB of RAM, 1 TB > Memory is more important than processor speed. > Would you really run 5000 users on one server? No.
- 14. Case Study 14 > eZuce > 5000 Concurrent User System
- 15. eZuce’s Corporate System 15 > Runs in AWS > m1.medium (1 core, 2 Amazon Compute Units, 3.7 GB RAM, 80 GB Storage) > 55 – 65 Concurrent Registrations > Conference bridge good to ~ 22 Concurrent callers (mix of inside / outside)
- 16. eZuce’s System in AWS Cloud 16
- 17. Private Cloud Case Study 500 Users 17 > Proxy / Registrar / Config Server / SIP Capture Server (qty 1) > 8 GB of RAM, 4 Cores, 200 GB > Proxy / Registrar (qty 2) > 8 GB of RAM, 4 Cores, 100 GB > Voicemail / Conference Bridge (qty 2) > 8 GB of RAM, 4 – 6 Cores, 300 GB
- 18. Customer running in VMWare Private Cloud 18