private cloud deployment model for academic institution · private cloud deployment model for...

Private Cloud Deployment Model for Academic Institution

Thinn Thu Naing, Ph.D.

Pro-Rector University of Computer Studies, Taunggyi

10h December 2015

MYANMAR

MYANMAR

University of Computer Studies, Taunggyi

Cloud Infrastructure

Services Available on the Cloud

Back-End Data Center (Cloud Service Provider)

User Developer Developer

Software as a Service

Platform as a Service

Infrastructure as a Service

Data Storage as a Service

Academic Cloud Deployment Model (Previous Work)

Open Source Systems involved in Private Cloud System Deployment

(Infrastructure as a Service (IaaS))

(Data Storage and Parallel Data Processing Framework)

(Platform of Google App Engine implementation framework)

(Moodle Learning Management System (LMS) Application on Cloud Infrastructure)

Open Source System for Private Cloud Infrastructure

Elastic Utility Computing Architecture Linking Your Programs To Useful Systems

Its origins are as a research project in the Computer Science department at the university of California, Santa

Barbara (UCSB)

Developer(s) Eucalyptus Systems, Inc. Initial release 1.0 2008-05-29

Stable release 3.2.0 (December 19, 2012)

Written in Java, C Operating system

Linux, can host Linux and Windows VMs

Platform Hypervisors (Xen, KVM, VMware)

Type Private and hybrid cloud computing

License GPLv3 (only), with Proprietary relicensing

Front- End Nodes Specification

Hardware Minimum Suggested Notes

CPU 1 GHz 2 x 2GHz For your cloud to scale it helps to have at least a dual core processor

Memory 512MB 2 GB Block storage serving benefits from the presence of memory for caching

Disk 5400rpm IDE 10000rpm SATA Block storage serving benefits from fast disk access

Disk Space 40GB 1TB Disk space will condition block stora capacity

2 Network interfaces 100Mbps 1000Mbps

Cluster will be subject to heavy traffic on a busy cloud setup as it will act as a router for all instances started on NC it Controls and will provide block storage (if needed) to them

Back-End Node Specifications Hardware Minimum Suggested Notes

CPU VT extensions

VT, 64-bit, Multicore

64-bit can run both i386, and amd64 instances; by default, Eucalyptus will only run 1 VM per CPU core on a Node

Memory 1GB 4GB additional memory means more, and larger guests

Disk 5400rpm IDE 7200rpm SATA or SCSI

Eucalyptus nodes are disk-intensive; I/O wait will likely be the performance bottleneck

Disk Space 40GB 100GB images will be cached locally, Eucalyptus does not like to run out of disk space

Networking 100Mbps 1000Mbps machine images are hundreds of MB, and need to be copied over the network to nodes

Physical Components of Private Cloud System with Eucalyptus

Front-end Servers Layer

Back-end Servers Layer

Virtualized Servers Layer

Type of VMs instances in Eucalyptus Private Cloud

• Instruction Set VMs (Amazon EC2, Eucalyptus) – Ubuntu VMs

CentOS VMs

– Windows Server 2003 VMs

• Framework VM

– AppScale VMs

15

? • Google’s MapReduce inspired Yahoo’s Hadoop. • Distributed large data computing framework

– For clusters of computers – Thousands of Compute Nodes

• Fault tolerant and scalable storage of very large datasets across machines in a cluster.

• Now part of Apache group • Consists of two components

– File Store (Hadoop Distributed File System(HDFS)) – A Distributed Processing System (Map/Reduce

Model)

Hadoop Storage System with PCs

• Specification • Name node

– 4GB (RAM) – 320 GB (HDD) – 2 Cores CPU

• Data node – 128 MB (RAM)

or low – 80 GB (HDD) or

low

Google App Eingine (GAE) Overview

• Write user application in Python or Java • Test locally • Deploy on Google public cloud infrastructure • Automatic Scaling • Pay-as-you-go

Free for limited quotas Pay for additional scale: CPU, Bandwidth, Data Storage

AppScale

• Open source GAE Implementation framework on private cloud

• Distributed and scalable API implementations • Infrastructure

– KVM/Xen – Eucalyptus – Amazon EC2

• Design and implement by UCSB

Architecture of AppScale Cloud and Hadoop Storage System Integration

Moodle System

Moodle Servers in Different Locations

Virtualized Moodle Servers on Private Cloud

Academic Cloud Deployment Model with CloudStack

THE PROJECT FOR DEPLOYING PRIVATE CLOUD SYSTEM FOR VIRTUALIZED LABORATORYS FOR

ICT HIGHER EDUCATION

Project Description

Title HRD Programme for Exchange of ICT Researchers and Engineers 2013

Granter Organization

APT (Asia Pacific Telecommunity)

Partner Organizations

Waseda University and KDDI Foundation, Japan

Period June 2014 to April 2015

Private Cloud System

• Private Cloud Computing Lab underlying University Network backbone

Logical Design of Private Cloud System

Design Consideration for Cloud Deployment Model

CloudStack Servers Specification Computer Node and Primary

Storage Server Secondary Storage Server

Brand PowerEdge R920 Brand Synology NAS (8 Bays)

Professor 2 CPU x Intel Xeon E7-4809 v2 Processor 1.9GHz, 12M Cache, 6.4 GT/s QPI, No Turbo, 6 Core, 105W

Professor Dual Core 2.13GHz CPU

Memory 128GB (8 X 16GB RDIMM, 1600 MHz, Low Volt, Dual Rank, x4)

Memory 2 x 4 GB DDR3 RAM

HDD 10Nos.x 1TB 7.2K RPM, 6Gbps Near Line SAS 2.5" Hard Drive - HotPlug

HDD 4 x Western Digital 2 TB ( Red ) Hard Disk

Controller PERC H730P Adapter RAID Controller, 2GB NV Cache

Software Design and Deployment Plan Server

Component

Operating System and Softwares Remarks

Management Server CloudDB Server Compute Node

Operating System : CentOS 6.5 x86 Other Server : MySQL Server Cloud Server : CloudStack version 4.4.1 Virtualization : KVM hypervisor Server

(1) Configure the network

(2) Host name (3) SELinux (4) NTP (5) Configuring the

CloudStack Package Repository

Primary Storage Server Secondary Storage Server

Operating System : CentOS 6.5 x86 File Server : NFS Server (NFSv4)

Implementation

• Phase I (Installation) • Install CloudStack components on the back end servers. • Provide IaaS (Infrastructure as a Service) and virtualized

computing resources including virtual machines, virtual switches, and virtual storage and so on.

• Phase 2(Testing) • To establish virtualized labs which can be involved to

provide computing services • Phase 3 (Implementing and Evaluation)

– This phase will evaluate the efficiency and effectiveness of academic cloud system and virtualized labs.

Disaster in Myanmar (Cyclone Nargis in 2008)

Disaster in Myanmar (Flooding in Myanmar 2015)

Thank you very much