Date post: | 01-Sep-2014 |
Category: |
Technology |
Author: | crs4-research-center-in-sardinia |
View: | 749 times |
Download: | 0 times |
Tecnologie Cloud per la fornitura di applicazioni
Copyright NICE - 2012
Tecnologie Cloud per la fornitura di applicazioni ambientali HPC in modalità SaaS
Paolo Maggi <[email protected]>R&D Manager
Tecnologie Cloud per la fornitura di applicazioni
Copyright NICE - 2012
Tecnologie Cloud per la fornitura di applicazioni ambientali HPC in modalità SaaS
High PerformanceComputing
SoftwareasaService
What is High Performance Computing?
� Refers to the use of supercomputers or clusters of computers to solve difficult computational problems that typically arise through scientific inquiry.
� Computer users turn to HPC when a problem is too large to solve on a conventional laptop or workstation or runs too slowly
Copyright NICE - 2012
runs too slowly– it requires too much memory or disk space– the algorithm is complex– the dataset is large– data access is slow
When we need HPC?
� To do a time-consuming operation in less time– I am an aircraft engineer– I need to run a simulation to test the stability of the wings at high
speed– I’d rather have the result in 5 minutes than in 5 days so that I
can complete the aircraft final design sooner
Copyright NICE - 2012
When we need HPC?
� To do an operation before a tighter deadline– I am a weather prediction agency– I am getting input from weather stations/sensors– I’d like to make the forecast for tomorrow before tomorrow
Copyright NICE - 2012
When we need HPC?
� To do a high number of operations per seconds– I am an engineer of ACME.com– My Web server gets 1,000 hits per seconds– I’d like my web server and my databases to handle 1,000
transactions per seconds so that customers do not experience bad delays
Copyright NICE - 2012
How to get my answer faster?
� Work harder– Get faster hardware
� Work smarter– Use optimized algorithms (libraries!)– Write faster code (adapt to match hardware)
Copyright NICE - 2012
– Write faster code (adapt to match hardware)– Trade convenience for performance (e.g. compiled program vs.
script program)
� Delegate parts of the work– Parallelize code– Use accelerators (GPGPU/CUDA)
Leveraging parallelism
Copyright NICE - 2012
Leveraging parallelism
Copyright NICE - 2012
Functional parallelismDifferent people are performing different tasks at the same time
Leveraging parallelism
Copyright NICE - 2012
Data parallelismDifferent people are performing the same task, but on different (equivalent) objects
What a cluster is…
� A cluster needs:– Several computers, nodes , often in special cases for easy
mounting in a rack– One or more networks (interconnections ) to hook the nodes
together– Software that allows the nodes to communicate with each other
(e.g. MPI)
Copyright NICE - 2012
(e.g. MPI)– Software, called Distributed Resource Manager (DRM), that
manage resources and assign them to individual users and computational jobs (e.g. LSF, PBS, SGE, …)
� A cluster is: all of those components working together to form one big computer (to give you answers in a fast way)
How a cluster looks like?
Copyright NICE - 2012
What is Cloud Computing?
Copyright NICE - 2012
A working definition of Cloud Computing ( NIST)
� Cloud computing is a model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications , and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.
Copyright NICE - 2012
management effort or service provider interaction. � This cloud model promotes availability� Composed of
– five essential characteristics– three service models– four deployment models
The NIST Cloud definition framework
Service Software as a Platform as a Infrastructure as a
EssentialCharacteristics
Resource Pooling
Broad Network Access Rapid Elasticity
Measured Service
On Demand Self-Service
Copyright NICE - 2012
CommunityCommunityCloudCloud
Private Private CloudCloud
Public CloudPublic Cloud
Hybrid Clouds
DeploymentModels
ServiceModels
Software as a Service (SaaS)
Platform as a Service (PaaS)
Infrastructure as a Service (IaaS)
Cloud Computing Offerings
Copyright NICE - 2012
About NICE
� Pioneer in Technical and Engineering Cloud solutions� 14 years experience with enterprise
Grid/Cloud solutions throughout all industries
� Worked closely with ISV and HW vendors since the beginning
� Strong relationship with Research and Academia
Headquarters: Italy
Offices: USA, Germany
Copyright NICE - 2012
� Core business: Access to Grid / HPC solutions and Remote Visualization� EnginFrame Grid Portal product line
� DCV Remote Visualization technology
� Other relevant competencies � Distributed Resource Management
� Grid Intelligence
� Visualization technology integration
NICE Customers and Market Segments
Automotive & Industrial Equipment
Audi, ARRK, Bridgestone, Bosch, Daimler, Delphi, Dow Chemical, FIAT, Ferrari, Honda,
Energy
Addax Petroleum, AECL, Hess, Bayerngas, BHP Billiton, Beicip, British Gas, Centrica, Chevron, Conoco-Phillips, Dong, Dowell, DSC-Libya, ENI/Agip, GazPromNeft, GDF, Logelco, Maersk Oil, Marathon Oil, Nexen, Novatek, Papuan Oil, Rosneft, Schlumberger, Shengli Oil, Sonatrach, Statoil, Talisman Energy, TNK-BP, TNNC, TOTAL, WesternGeco, Xinjiang Oil
Aerospace & Manufacturing
AIRBUS, Air Products and Chemicals, Procter&Gamble, SelexGalileo, Goodrich Aerospace, Hamilton Sunstrand, Kimberly Clarke, Magellan Aerospace, Nordam, Northrop Grumman, Raytheon, Thales
Copyright NICE - 2012
Delphi, Dow Chemical, FIAT, Ferrari, Honda, Hyundai, Jaguar-Land Rover, Lear, Magneti Marelli, McLaren, P+Z, PSA, Tata Steel, Toyota, TRW, Volkswagen
Life Sciences
Bayer, Biolab, DEISA project, Swiss Institute for Bioinformatics, Partners Healthcare, Pharsight, M.D. Anderson Cancer Center
TOTAL, WesternGeco, Xinjiang Oil
Research & Education
ASSC, Beihang Uni, CCLRC, CILEA, CNR, CNRS/IN2P3, ENEA, Georgia State Uni, INFN, RMSC, Harvard Uni, Messina Uni, Huazhong Normal Uni, Yale Uni
Others
Accent, Samsung SDI, SensorDynamics, Bank of Italy, Deutsche Bank
The NICE Grand Vision
InternalResources
Administrators& Managers
Supplier
Partner
Copyright NICE - 2012
Customer
HPC Centers, Public Clouds,
…
User
Multi-Tenancy &Collaboration services
NICE Products & Solutions
�� NICE EnginFrameNICE EnginFrame
�� NICE DCVNICE DCV
Copyright NICE - 2012
�� NICE EnginFrame ViewsNICE EnginFrame Views
�� Solutions for vertical marketsSolutions for vertical markets:
� Life Science
� Oil&Gas
� Automotive
NICE Products Overview
Service Service CC
((SaaSSaaS))
Service Service CC
((SaaSSaaS))
Service A Service A
(HPC)(HPC)
Service A Service A
(HPC)(HPC)Service Service XX
((workflowworkflow))
Service Service XX
((workflowworkflow)). . .Service BService B
(3D application)(3D application)
Service BService B
(3D application)(3D application)
Customer 1 (Corporate network) Customer N. . .Customer 2 (VPN)
Vertical solutions
Service Presentation Layer(Web, API, Command-line, ...)
Project Mgr End Users Developers End UsersDevelopers End Users
Access ControlAD, LDAP, NIS, …
Copyright NICE - 2012
GridMLdriver
CloudMLdriver
ViewsMLdriver
VMWARE, EC2, MOAB, ISF, …
DCV, VNC, RGS, TAW, …
LSF, Lava, SGE, PBS, MOAB, …
Infrastructure model abstraction
Desktop CloudVisualization
Vertical solutions
DataGateSLB-RE
…HPC
Views
Accounting & Billing
RDBMS
Linux & Windows
3D sessions
Our architecture
Collaborators,Support staff
Self-Service Offering
End Users
Thin viewer
ResourcesAccess
VisualizationServers
HTTP(S)H
PC
schdulers
DCV protocol
Copyright NICE - 2012
End Users schdulers
HPC jobs
Command-line
Developers, Integrators
SOAP
Orchestration &
P
rovisioning
Storage
User Experience in the Technical CloudOpen Grid
Portal
CommercialHPC Cloud
HPC
Enterprise Grid
Desktop Scavenging
Copyright NICE - 2012
HPC Clusters
NICE EnginFrame key features
� Lightweight Web front-end for HPC systems� Best in class data management� WebServices API and command-line interface� Controlled, user friendly application submission
Supports all mainstream job schedulers
Copyright NICE - 2012
� Supports all mainstream job schedulers� Workload, resource, license monitoring
Batch job / workflow submission
User friendly,Application-oriented
Job submission
Copyright NICE - 2012
Flexible and efficientInput file management
Hide complexity ofUnderlying scheduler
Application Data Management
Application data can be organized into projects
Application data can be marked as starred
Copyright NICE - 2012
starred
Metadata can be associated to application data
Monitoring
Jobs, Hosts, Queues, Licenses, …
Copyright NICE - 2012
Data transfers & file management
The file manager component allows to seamless navigate and access server-side files from the web browser
Copyright NICE - 2012
SOA-enabled job submission
� WS-I interface� Java / .NET client library and
command line interface� Simplifies integration with
client-side applications (optimization, workflow, etc.) for power-users
Copyright NICE - 2012
What is a 3D Cloud?
� "Cloud computing is a model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications , and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.” (NIST)
Copyright NICE - 2012
management effort or service provider interaction.” (NIST)
� 3D Clouds will enable on-demand network access to interactive 3D applications (like visualization applications for scientific data, CAD applications, etc.)
Why NICE customers need a 3D Cloud?
A typical scenario in the CAE computing sector
1 - Simulation submission
Win
2 - Data Transfer to visualize results
Copyright NICE - 2012
Server/Data Center
Win
2 - Data Transfer to visualize results
3 - CollaborationNeeds Data exchange CAE Workstations
• Can run small to medium serial analysis• Requires high-end GPU for rendering• May represent IT management
challenges (tens to hundreds of seats)
Problems associated to this model
NETWORKNetwork overload leading to poor performance and response times all round
COSTExpensive, dedicated workstations (GPU, memory, …) with short lifecycle
IT MANAGEMENTSupport, update and replace tens to tens of thousands of workstations
Copyright NICE - 2012
WORKSTATION SIZINGWorkstations have to be sized for the largest expected models
SECURITYMoving sensitive data around (in/out organization) is always risky
WORKFORCECurrent models do not support a diverse mobile workforce
Collaboration
Moving technical applications closer to the data
PCWS
HPC job
submission and
management
HPC job
submission and
management
Remote
Visualization
Remote
Visualization
Copyright NICE - 2012 33
Linux
Job submission
managementmanagement VisualizationVisualization
Benefits of the HPC Cloud model
NETWORKNetwork is no more a bottleneck and data loads faster
COSTCentralized & Shared servers are less expensive to buy & manage
IT MANAGEMENTSupport, update and replacement are more efficient & do not affect users
Copyright NICE - 2012
WORKSTATION SIZINGResources are dynamically sized based on users needs
SECURITYSensitive data remain within protected data center with full access control
WORKFORCEUsers can virtually connect & collaborate from anywhere with any client
� Technologies providing secure remote network access to interactive 3D applications leveraging server-side graphic hardware acceleration (GPUs)
Enabling Technologies for the 3D Clouds
Copyright NICE - 2012
� A software stack that allows the end-user to easily launch and access remote interactive applications and takes care of managing and load balancing applications and desktop sessions running within a Visualization Farm
NICE EnginFrame Views
RealVNC
2D & 3D Applications2D & 3D Applications
Job Scheduler & Distributed Resource Manager
NICE EnginFrame (Portal and Gateway)
VirtualGL NICE DCV HP RGS NX, RDP, ...
Copyright NICE - 2012
RealVNC
Storage Infrastructure
VirtualGL NICE DCV HP RGS
Computing Infrastructure
LINUX
SOLARIS
WINDOWS
NX, RDP, ...
How Does It Work?
EnginFrame
1
The user connects to EnginFrame to create a new Session
2
Job sch. selects the appropriate node and starts the session
3
User gets connected to the session
Win
Copyright NICE - 2012
Job Sch.
64GBytes, Linux
8GB, Windows
4GB, Win/Lin blades
User gets connected to the session on the selected node
Heterogeneous infrastructure: HW, OS, middleHeterogeneous infrastructure: HW, OS, middle--wareswares
Benefits for the End Users
� Access to Applications as a Service– Hide infrastructure details (site, platform, etc…)– Session allocation can be influenced by memory requirements, data
location affinity and other customer-specific parameters– Intelligent load balancing of sessions, based on the Job Scheduler
• Memory-aware, Memory reservation, Application license-aware– Fewer or better optimized data transfers
� Collaboration, session sharing– The session owner can generate a “URL” that can be sent by email /
Copyright NICE - 2012
– The session owner can generate a “URL” that can be sent by email / instant message to invite a collegue to join a given session, without disclosing the user’s password
� Easy management of active session– Create new sessions with user-specific preferences (resolution, etc...)– List , reconnect and kill existing sessions
� Seamless Access to the sessions and Single Sign-On– Automation of session-level password create/destroy
• E.g. login via NTLM / ActiveDirectory credentials and map to Linux user
Benefits for the Administrators
� Increased level of service provided to users– Sessions are load balanced by the Job Scheduler to match user needs– Memory reservation and Data locality scheduling– Reduce help desk calls– Exposed services can be personalized per user/group/project
� Accounting– Sessions are jobs, so the resource usage accounting by user, group,
project can be collected through any Analytics toolMonitoring
Copyright NICE - 2012
� Monitoring– The load and usage of the login farm is monitored via EnginFrame– Node loading conditions, active sessions– Administrators can control and manage users’ idle or stuck sessions
� Support– Administrators can connect to user’s sessions to provide support
� Security– Easy integration into identity services, SSO, Enterprise portals
NICE DCV (Desktop Cloud Visualization)
� Originated in IBM Research in 2004, acquired by NICE in summer 2010
� Provides efficient remote access to graphic-intensive, professional OpenGL applications
� A central (graphics-enabled) server is accessed by remote users
Copyright NICE - 2012
� Users only need low-end machines with network connectivity to the server in order to view and interact with remote application
� One or more users (collaborators) can simultaneously access the server
� Is the key technology for delivery of real-time 3D graphics but relies on other software to provide the collaborative environment (RealVNC Visualization Edition)
NICE DCV – Main Features
� Works on Windows and Linux, leveraging high-end NVIDIA GPUs
� Support GPU sharing across multiple users, multiple OS
– First product on the market to allow the sharing of physical GPUs between multiple Windows VMs while maintaining full OpenGL acceleration and workstation-class performance
� Bandwidth optimized balancing of quality Vs. frame rate
Copyright NICE - 2012
� Bandwidth optimized balancing of quality Vs. frame rate
� High network latency tolerance
� Validated and optimized for Technical Computing applications
� Perfectly integrated with NICE EnginFrame for session management
Windows application delivery
Copyright NICE - 2012
Linux application delivery
43
Copyright NICE - 2012
Web-based Thin client access
Copyright NICE - 2012
Built-in collaboration45
Copyright NICE - 2012
Thank you!
Copyright NICE - 2012