Transcript Health-e-Child Gateway

Health-e-Child: An Integrated
Platform for European Paediatrics
Konstantin Skaburskas
CERN IT/GD
07 Sept 2007, CoreGRID Summer School
Budapest
Project Objectives
• Establish Horizontal and Vertical integration of data, information and knowledge
• Develop a Grid-based biomedical information platform, supported by sophisticated and
robust search, optimisation, and matching techniques for heterogeneous information,
• Build enabling tools and services that improve the quality of care and reduce its cost by
increasing efficiency
• Integrated disease models exploiting all available information levels
•
Database-guided decision support systems
•
Large-scale, cross-modality information fusion and data mining for knowledge discovery
• A Knowledge Repository for Paediatrics
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Project General Info
 Instrument:
Integrated Project (IP) of the
Framework Program FP6
 Project Identifier: IST-2004-027749





Coordinator:
Partner:
Timetable:
Total cost:
EC funding:
 Web page:
Health-e-Child
Siemens AG, Dr. Jörg Freund
14 European (companies, hospitals, institutions)
01-Jan-06 to 31-Dec-09 (4 years)
16.7 Mio. €
12.2 Mio. €
http://www.Health-e-Child.org
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Introduction
A Geographically Distributed Environment
ASPER
+
UCL
GOSH
Clinical Site
UWE
+
SIEMENS
CERN
R&D Site
An Integration
Challenge
+
NECKER
IGG
+
FGG
EGF
UOA
INRIA
MAAT
Health-e-Child
LYNKEUS
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
HeC System Overview
Cardiology
NeuroOncology
Rheumatology
Knowledge
Discovery
Decision Support
Systems
Disease Models
Common Client Applications
Ontological
Layer
Query Processing
Data
Management
Health-e-Child Gateway
Grid Middleware
IGG
Health-e-Child
Necker
GOSH
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Challenge
Data Integration Challenge (1)
IGG
Cardiology
+
NECKER
MS ACCESS + Excel
TOMCAT
NO - Paper-based
PACS
YES - But not operational
YES
NO
NO - Paper-based
NO - Paper-based
YES
NO
PACS
NO - PACS in 2007
- Several
DBMSs: MS Access, RADOS,
RIS, TOMCAT RADOS
…
DB
RIS
Not Available
- Potential
problems
accessing proprietary
systems … YES - But being tested
PACS
YES - But not operational
- Heterogeneous data, structured, unstructured …
Radiology
Molecular Genetics
NeuroOncology
Proteomics
GOSH
DB
Highlights
DB
MS ACCESS + Excel
Rheumatology
+
DB
MS ACCESS + Excel
PACS
NO
DB
MS ACCESS + Excel
PACS
NO
DB
MS ACCESS + Excel
PACS
NO
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
+
Challenge
Data Integration Challenge (2)
MRI @ IGG
X-Ray @ NECKER
US @ GOSH
Highlights
X-Ray @ IGG
- Several Imaging Modalities: CT, MRI, US …
PACS @GE,
GOSH
- Several Imaging Devices: SIEMENS,
HP, PHILIPS …
(rough estimation gives: 540’000 images)
CT @ NECKER
X-Ray @ GOSH
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Challenge
Applications Integration Challenge (1)
NeuroOncology
Cardiology
Rheumatology
Viewing remote cases
MRI, CT
MRI, US
MRI, US, (x-ray)
Distributed similarity search
Highlights
(involving image features)
- Similar size, location, and
enhancement of tumor  search
can be run on extracted features
- Similar extend of RV
or LV dilation, similar
shape
- (images are evaluated
by using various scores)
- Different Types of Applications: Image Processing, Data Mining,
Intensive image processing
- Segmentation and
- Training of model for GUI Clients, Services
…
DSS
registration with atlas
- Different Levels
of Integration:
standalone application, online
- Tumor
growth simulation
“gridified”
application,
external
services
… - Automatic
Interactive/automated
training or
- Interactive
refinement
(by
refinement of models
Algorithm assessment
(same algorithm with different
parameters and/or various
algorithms)
Health-e-Child
adjusting parameters)
incremental
refinement of the
model
-
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Challenge
Applications Integration Challenge (2)
IGG
Highlights
- Different Networks: LANs, WANs, Internet
- Security Constraints: Local & National Regulations
- Bandwidth Limitations: LAN/WAN & Internet uplinks …
NECKER
GOSH
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Challenge
Problem Description
Security
- Interconnecting Hospitals through the Internet
- Providing Services & Data Online
- Connecting a Research Platform to Hospitals’ Information Systems
- Manipulating Patient Data
Heterogeneity
- Dealing with several imaging devices, technologies, protocols, data, environments…
Distribution, Scalability
- Coping with geographically distributed sites across Europe (France, UK, Italy…)
- Building a scalable system for a future expansion
Solution
Privacy
Usability
- End-users: Clinicians
- Providing a rapid and efficient access to the Platform
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Challenge
Problem Description
Security
- Interconnecting Hospitals through the Internet
- Providing Services & Data Online
- Connecting a Research Platform to Hospitals’ Information Systems
- Manipulating Patient Data
R
C
N
H imaging devices, technologies,
A
E
- Dealing with several
protocols, data, environments…
E
R
U
U
D
R
Distribution, Scalability
M
I
O
- Coping with geographically
distributed sites across
Europe (France, UK, Italy…)
- Building a scalableAsystem for a future expansion
O
Heterogeneity
Solution
Privacy
Usability
- End-users: Clinicians
- Providing a rapid and efficient access to the Platform
Horizontal Platform
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Challenge
Data Flow
R
H
E
U
M
A
C
A
R
D
I
O
N
E
U
R
O
GOSH
NECKER
Horizontal Platform
IGG
+
+
Health-e-Child
+
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Solution
Our Approach
Highlights
- Simplicity, Abstraction from the complexity of Grid
- Modularity & Scalability, Off-the-shelf components
- State-of-the-art Approaches
One Key
Access
Point
+
To enter
the system Hospital X
Per
Institution
User
Workstation
Health-e-Child
Gateway
+
+
+
+
+
+
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Solution
Our Approach
+
+
Access
Point
+
Hospital X
Per
Institution
User
Workstation
Health-e-Child
Gateway
+
+
+
+
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
The Platform
HeC
Scheduling
Info
Monitoring
System
Hosting Domain 1 Stable & Secure Environment
EGEE gLite
Health-e-Child
HeC Gateway
HeC
DBMS
Storage
200GB
50GB
1TB
Security
&
Registration
Data Unit
Access
Point
Job
Managnt
Computation Unit
The Health-e-Child Access Point
Hosting Domain 2 Stable & Secure Environment
Virtualization…
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
The Platform
The Health-e-Child Gateway (1)
Client Applications
Access
Point
Functionality Access
HeC Gateway
Infrastructure Abstraction
Hospitals’ IS
Knowledge
Computing
gLite InfraResources
Repository
Inside the box…
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
The Platform
The Health-e-Child Gateway (2)
Client
Client Applications
Applications
Access
Point
Ongoing
Portal
Service &
Ongoing
Functionality
Access Ongoing
Factory
Data
Integration
HeC Gateway
Authen. &
Author.
Services
Ongoing
Infrastructure
Abstraction
Grid Abstraction
DB Abstraction
gLite API
OGSA-DAI
AMGA
GT4*
Hospitals’ IS
gLite Infra
* GT4: Globus Toolkit 4 (WSRF Container)
* OGSA-DAI or AMGA: Data Access & Integration
Health-e-Child
Knowledge
Repository
Inside the box…
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Client Applications
Access
Point
Functionality Access
Health-e-Child SOA
HeC Gateway
Infrastructure Abstraction
Computing Resources
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Access
Point
Security
Services
Discovery &
Connectivity
Composition
HeC Gateway
Medical Query
Processing
Data
Integration
Health-e-Child
Grid
Data Distribution &
Federation
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Our Approach
+
+
+
One Key
+
+
To enter
the system Hospital X
+
User
Workstation
Health-e-Child
Gateway
+
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Security
Authentication
Access
Point
Client
Workstation
2. Authentication request
9. Check VO membership
Authentication
- Leveraged Security Model (2-FA is respected)
- HeC CA, Affordable Solution: both in terms of cost & effort
8. VOMS Request
- Gives us the freedom to make several tests
VOMS
10. Returns atributes blob
(since based our own CA)
Resource
- Certificates are used in different contexts
(e.g. logging)
Note:
- Tunnels: HTTPS + GSI Secure Conversation
- Heartbeat = resource & proxy lifetime extension
11. Store VOMS Proxy
1. Generate Grid Proxy
Authentication
Factory
3. Instantiates
Highlights:
4. Returns Endpoint
HeC Gateway
Health-e-Child
Grid Services
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Security
Logging
•
Logging
•
Log4j Appender
•
Extended Appender for fetching more detailed information
–
•
Portal
Almost no impact on perform times (much less than log in files)
Web Portal only available at the Institution
•
•
•
UserID, Layer, Service, Operation, Log Level, Timing …
Appender is asynchronous
–
•
Auth
Log Database
Misc
Access is restricted to server administrator (HTTPS)
Uses HeC Certificates
HeC Gateway
Functionality
•
•
Log can be sorted by user/layer/service/operation…
Statistics can be processed
–
–
–
–
–
•
Services/Layer reliability,
Services/Layer load,
User Performed Operations,
Platform Usage,
Layer/Service/Operation & Platform Performance
Log Traces can be generated
–
Possibility to export traces as text file
Web Portal
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Access
Point
Security
Services
Discovery &
Connectivity
Composition
HeC Gateway
Medical Query
Processing
Data
Integration
Health-e-Child
Grid
Data Distribution &
Federation
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Connectivity
Client Connectivity (1)
Client Type2
P2P3*
HTTPS
Stubs
Stubs
Stubs
ServiceLibs
GW Stack
Stubs
ServiceLibs
GW Stack
GW Stack
ServiceLibs
Client Base
ServiceLibs
Client Base
Client Type3
P2P3 Stack
Client Type1
HTTPS + XML Conversation
HeC Gateway
Grid Infrastructure
*P2P3: Peer-To-Peer Patient Privacy
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Access
Point
Security
Services
Discovery &
Connectivity
Composition
HeC Gateway
Medical Query
Processing
Data
Integration
Health-e-Child
Grid
Data Distribution &
Federation
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Data Integration
DICOM Files
Access
Point
Client
Workstation
HEARTBEATS
Authentication
HEARTBEATS
Portal Factory
Highlights:
CE
6. Store/Register
.5. Pick-up
gridFTP
7. Thumbnails
2. Fire-up
- Asynchronous operations
1. Portal Call (START TRANSF)
- Thumbnails generated
on grid from DICOM
Portal slices
4. Portal Call (ADD)
(useful for efficiently browsing patient records)
Resource
- DICOM Compliant
Notification
SE
LFC
Tmp
3. File Transfer
Health-e-Child
HeC Gateway
Grid Services
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Access
Point
Security
Services
Discovery &
Connectivity
Composition
HeC Gateway
Medical Query
Processing
Data
Integration
Health-e-Child
Grid
Data Distribution &
Federation
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Grid Middleware
Grid Infrastructure
GOSH
UWE
SIEMENS
CERN
NECKER
IGG
UOA
MAAT
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007
THANK YOU FOR YOUR ATTENTION
www.health-e-child.org
Health-e-Child
Konstantin Skaburskas – CoreGRID Summer School, Budapest, 07 Sept 2007