Solution for private cloud computing



Similar documents
Parallels Plesk Automation

Module I-7410 Advanced Linux FS-11 Part1: Virtualization with KVM

How To Run A Cloud Server On A Server Farm (Cloud)

Virtualization Management the ovirt way

2) Xen Hypervisor 3) UEC

Shoal: IaaS Cloud Cache Publisher

Deploying Business Virtual Appliances on Open Source Cloud Computing

This presentation covers virtual application shared services supplied with IBM Workload Deployer version 3.1.

IBM Cloud Manager with OpenStack

ovirt Introduction James Rankin Product Manager Red Hat Virtualization Management the ovirt way

OpenNebula Open Souce Solution for DC Virtualization

Virtual Machine Management with OpenNebula in the RESERVOIR project

Regional SEE-GRID-SCI Training for Site Administrators Institute of Physics Belgrade March 5-6, 2009

OpenNebula Open Souce Solution for DC Virtualization

Cloud Computing for Control Systems CERN Openlab Summer Student Program 9/9/2011 ARSALAAN AHMED SHAIKH

SUSE Cloud 2.0. Pete Chadwick. Douglas Jarvis. Senior Product Manager Product Marketing Manager

RingStor User Manual. Version 2.1 Last Update on September 17th, RingStor, Inc. 197 Route 18 South, Ste 3000 East Brunswick, NJ

INTRODUCTION TO CLOUD MANAGEMENT

FleSSR Project: Installing Eucalyptus Open Source Cloud Solution at Oxford e- Research Centre

EZblue BusinessServer The All - In - One Server For Your Home And Business

OpenNebula Open Souce Solution for DC Virtualization. C12G Labs. Online Webinar

RED HAT ENTERPRISE VIRTUALIZATION

Integration of Virtualized Workernodes in Batch Queueing Systems The ViBatch Concept

OGF25/EGEE User Forum Catania, Italy 2 March 2009

Backup & Disaster Recovery Appliance User Guide

Sistemi Operativi e Reti. Cloud Computing

Building Storage Service in a Private Cloud

SolidFire SF3010 All-SSD storage system with Citrix CloudPlatform Reference Architecture

Ezilla - WebOS Toward the Private Cloud & Possibility of Virtual Classroom

NOC PS manual. Copyright Maxnet All rights reserved. Page 1/45 NOC-PS Manuel EN version 1.3

Chapter 11 Cloud Application Development

Very Large Enterprise Network, Deployment, Users

OnApp Cloud. The complete platform for cloud service providers. 114 Cores. 286 Cores / 400 Cores

METAARCHIVE & CLOUD COMPUTING

SURFsara HPC Cloud Workshop

Automated deployment of virtualization-based research models of distributed computer systems

Assignment # 1 (Cloud Computing Security)

THE EUCALYPTUS OPEN-SOURCE PRIVATE CLOUD

Ryu SDN Framework What weʼ ve learned Where weʼ ll go

A quantitative comparison between xen and kvm

IBM Platform Computing Cloud Service Ready to use Platform LSF & Symphony clusters in the SoftLayer cloud

RED HAT ENTERPRISE VIRTUALIZATION & CLOUD COMPUTING

Overview Customer Login Main Page VM Management Creation... 4 Editing a Virtual Machine... 6

Introduction to ovirt

Introduction to OpenStack

Proactively Secure Your Cloud Computing Platform

Linux Virtualization Nesting and Management

Using SUSE Cloud to Orchestrate Multiple Hypervisors and Storage at ADP

How to Backup and Restore a VM using Veeam

SEP Disaster Recovery and Backup Restore: Best

VMTurbo Operations Manager 4.5 Installing and Updating Operations Manager

OVERVIEW. The complete IaaS platform for service providers

Cloud UT. Pay-as-you-go computing explained

With Red Hat Enterprise Virtualization, you can: Take advantage of existing people skills and investments

EZblue BusinessServer The All - In - One Server For Your Home And Business

Description of Application

Marco Mantegazza WebSphere Client Technical Professional Team IBM Software Group. Virtualization and Cloud

Red Hat Network Satellite Management and automation of your Red Hat Enterprise Linux environment

Aspen Cloud Server Management Console

Making a Smooth Transition to a Hybrid Cloud with Microsoft Cloud OS

Linux/Open Source and Cloud computing Wim Coekaerts Senior Vice President, Linux and Virtualization Engineering

Red Hat Satellite Management and automation of your Red Hat Enterprise Linux environment

CERN Cloud Infrastructure. Cloud Networking

CTERA Cloud Onramp for IBM Tivoli Storage Manager

Definitions. Hardware Full virtualization Para virtualization Hosted hypervisor Type I hypervisor. Native (bare metal) hypervisor Type II hypervisor

Managing your Red Hat Enterprise Linux guests with RHN Satellite

The OpenNebula Cloud Platform for Data Center Virtualization

Very Large Enterprise Network Deployment, 25,000+ Users

Deployment of Private, Hybrid & Public Clouds with OpenNebula

Research of Enterprise Private Cloud Computing Platform Based on OpenStack. Abstract

Technology and Cost Considerations for Cloud Deployment: Amazon Elastic Compute Cloud (EC2) Case Study

Intel Cloud Builder Guide to Cloud Design and Deployment on Intel Xeon Processor-based Platforms

A Study on Architecture of Private Cloud Based on Virtual Technology

Cloud.com CloudStack Community Edition 2.1 Beta Installation Guide

1 Copyright 2011, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 7

BLACK BOX. Quick Start Guide. Virtual Central Management System (VCMS) Works with LES Series Console Servers. LES-VCMS. Customer Support Information

Open Source Cloud Computing Management with OpenNebula

StACC: St Andrews Cloud Computing Co laboratory. A Performance Comparison of Clouds. Amazon EC2 and Ubuntu Enterprise Cloud

Uptime Infrastructure Monitor. Installation Guide

Restricted Document. Pulsant Technical Specification

WebLogic on Oracle Database Appliance: Combining High Availability and Simplicity

Altor Virtual Network Security Analyzer v1.0 Installation Guide

Distributed Block-level Storage Management for OpenStack

How to Guide: StorageCraft Cloud Services VPN

Wazza s QuickStart 1. Leopard Server - Install & Configure DNS

RED HAT ENTERPRISE VIRTUALIZATION FOR SERVERS: COMPETITIVE FEATURES

Transcription:

The CC1 system Solution for private cloud computing 1

Outline What is CC1? Features Technical details Use cases By scientist By HEP experiment System requirements and installation How to get it? 2

What is CC1? The CC1 system is a complete solution to manage computing resources in the form of IaaS in any institution or company. It helps to organize and virtualize heterogeneous compute and storage resources. Good for small (100 cores) and large installations (>10000 cores). Final user can create and save virtual machines in easy way through the Web interface or EC2 compatible API. 3

Features For the end users: CC1 provide Web interface and EC2 programming interface which is easily accessible by Python scripts. Disk space for users is provided as a space for virtual machine images and virtual storage disks there is no user space like a cloud object storage. With CC1 system user gets management options for virtual machines (VM) and virtual farms, which are group of Vms forming a computing cluster with one head node. 4

Features For the CC1 administrator Dedicated administrator panel to manage users and resources. Administrator can organize scattered clusters unified with one CC1 cloud controller. Elastic way to extend/narrow number of physical nodes in the system. 5

Features: administrator module List of clusters added to CC1 cloud List of storages added to CC1 cloud Adding new storage Adding new cluster List of physical nodes for one of cluster 6

Features CC1 administrator can easily set and modify resources limits for particular users, also can set monthly credits to use resources. User can check usage of credits on chart. 7

Features User can monitor each VM 8

Technical details 9

Technical details 10

Technical details Software: Tested under latest stable Debian release. Uses virtualization package KVM (libvirt) and web application framework Django. New version is released every few months. Uniform Python environment Interface WWW (Django with Apache). CLM (Cloud Manager) and CM (Cluster Manager). Databases (PostgreSQL or MYSQL). Communication (more RESTful communication since version 2.0 instead of XML RPC). All CC1 services are recommended to be installed on separate virtual machines for easy backup and fast recovery. 11

Technical details Virtual Machines Libvirt is used as low level virtualization tool (not a bare metal solution). The KVM is default hypervisor. Libvirt supports more hypervisors, but were not tested. VM image is copied from physical storage to a local disk on the node and from this clone of image VM is started. It is recomended to keep the size of VM images small (10 GB for Linux). Additional space can be provided via virtual disks. Possibility to create preconfigured virtual computing clusters (Farms) with batch system. 12

Technical details Virtual storage disks User storage is provided via virtual disks that can be attached to VM at the start or to a running VM (USB and virtio drivers). Virtual disks are created by users on the physical storage through the network protocol. Disks are accessed directly from the physical storage. The disks are permanent i.e. survive after destruction of VM. 13

Technical details User's private networks Each user has its own network/networks. Range of private IP addresses may be booked by users from the total pool. Inter-VM communication is done by routing OSPF protocol (Quagga). Users virtual machines deployed on physical nodes 14

Technical details Network topology and administration CC1 Installation packages configure all necessary services required to start cluster's network. Network solution for VMs is based on small routed networks (not bridged). Each virtual network for VM: Is created using Libvirt. Is connected to physical node. Has a local Libvirt's dnsmasq (basic DNS and DHCP). All these networks are joined by routing protocols (OSPF). Private IP address is provided to VM by dnsmasq service (DHCP) managed by libvirt. Routing from the internal cluster network to the Internet is done on Routing server with proper quagga configuration. 15

Technical details Communication to VM from the outside VM Public IP can be dynamically attached to VM, then released and attached to another VM. Private to public IP mapping on the node where VM is running. VNC port mapping CC1 Cluster Manager forwards VNC ports to nodes, so needs to have its own public IP. 16

Use cases: by the Scientist Web interface allows to create of machine groups (so called farms). Scientist can process calculations on such pre configured computing cluster. For now, SLURM batch system is available. Easy to add other kinds of clusters. 17

Use cases: by HEP experiment HEP experiments like Belle II, LHCb, ATLAS, can use computing centers based on cloud computing paradigm. In particular, clouds with EC2 compatible interface (which is available in CC1 system). In the Belle II experiment, distributed environment is lead by the DIRAC system. DIRAC can handle jobs submitted to the grid and the cloud sites as well. 18

Use cases: by HEP experiment 19

Use cases: by HEP experiment 20

Use cases: by HEP experiment 21

Use cases: by HEP experiment 22

System requirements CC1 is a free Linux software build using modern virtualization and Internet technologies. CC1 System requirements: Full system can be installed on one powerful node. For bigger installations is recommended to split system to a few operational machines with the following parts: routing service, interfaces, cloud manager, cluster manager, databases. Router server to route packages between the Internet and cluster network (Quagga configuration) Hardware requirements: Physical nodes A few multi core nodes with hardware virtualization and Debian system installed Underlying host OS does not matter for end users Debian has good hardware support (including enterprise blade Systems) Disk storage Mounted to nodes as remote location (NFS by default, other distributed file systems also possible) For large installations disk array is recommended to have proper performance. Recommended network is 10 Gb/s. Optional: Mail server to sent activation request during user registration DNS to translate domain names for elastic (public) IP of VM. 23

System installation To install all CC1 components on the same machine Add CC1 repository to the operating system. Install deb packages: cc1-cm cc1-clm cc1-wi cc1-ec2 cc1-common Modify configuration files according to the installation instruction on http://cc1.ifj.edu.pl and restart services. Visit your machine address by Internet browser to see CC1 web portal. 24

Technical details CC1 installation at Institute of Nuclear Physics PAN in Krakow 1000 cores (Blades IBM HS22 Xeon L5640 @ 2.27 GHz) Disk storage 100 TB Network: 10 Gb/s external link 10 Gb/s backbone of internal network 25

How to get CC1? You can test the system ONLINE! Just send us request about access to CC1 Web and EC2 interfaces to cc1@cloud.ifj.edu.pl and we will provide a temporary access to computing resources in Krakow. CC1 project page is www.cc1.ifj.edu.pl There, in 'download' section you can find disk images with pre installed CC1 system. So you can test CC1 Web portal on your local machine. Do not hesitate to contact me! Rafal.Grzymkowski@ifj.edu.pl 26