Transcript CSE5810 - University of Connecticut

CSE 5810 Individual Research Project:
CSE
5810
Integration of Named Data Networking for
Improved Healthcare Data Handling
Robert Martin
Computer Science & Engineering Department
The University of Connecticut
371 Fairfield Road, Box U-255
Storrs, CT 06269-2155
[email protected]
Martin-1
Motivation
CSE
5810



Technology limitations
 Larger data files
 Multiple databases
 Ever expanding healthcare network
Fast pace hospital environment
 Providers constantly moving
 Intolerant to delayed data access
Lack of communication between departments
 Clinical, technical, business management,
financial, etc.
Martin-2
Large Scope
CSE
5810


Keep up with change
 Larger data files
 Mobile devices
 Real-time data availability
Conform to busy hospital environment
 Revamp current network
 Transparent infrastructure
Martin-3
Overall Goal
CSE
5810

Apply Named Data Networking within a hospital
environment
 Data connectivity
 Improved transmission speeds (compared with
regular IP networking)
 Improved mobile device handling
 Interoperability between diverse departments
Martin-4
Named Data Networking (NDN)
CSE
5810
Image adapted from:
Tsudik, Gene. NSF FIA PI meeting: “NDN team presentation.” Berkeley, CA, May 25, 2011.
Martin-5
Hospital Setting NDN
CSE
5810
Martin-6
NDN vs. IP Networking

CSE
5810

Named Data Networking
 Data centric approach
IP Networking
 Looks at where data is located
Image adapted from:
Jacobson et al. (full reference in notes)
Martin-7
Interest and Data Packets
CSE
5810

Interest Packet
 Data name in query
 Nonce is unique identifier
 Selectors help better match interest to data
 Scope and interest lifetime help guide
packet to intended data

Data Packet
 Content is of arbitrary data size
 Signature is used to verify the
packet’s producer and its integrity
throughout transmission
Martin-8
Pending Interest Table (PIT)
CSE
5810


Monitors all unsatisfied interest packets
Entry classified as unsatisfied until either a data packet
is received (to match its interest) or the interest
lifetime value is reached
“A Case for Stateful Forwarding Plane” by C. Yi et al. depicts a
great image for how node’s use PITs(see full reference in notes)
Martin-9
Forwarding Information Base (FIB)
CSE
5810

Monitors downstream data location through next hop
neighbor
“A Case for Stateful Forwarding Plane” by C. Yi et al. depicts a
great image for how node’s use FIBs(see full reference in notes)
Martin-10
Content Storage
CSE
5810




Cache data locally
Pushes data closer to consumer(s)
Allows network to become “data focused”
 Quicker fetching of data for consumer
Data architecture can vary
 FIFO, LRU, etc.
Martin-11
Data Naming
CSE
5810


Application specific
Flexible standards
 Classifications and standards can be adjusted
Martin-12
Security

Nurse fetching data which is unrelated to her role in
the hospital (e.g. Patent’s social security number)

Security integrated into data packet
 Authentication process
CSE
5810
Martin-13
Fetching Data

CSE
5810
Filtering naming system
 Adjust documentation standards for each
department
 E.g. Financial employee and patient see “heart attack”
vs. global view classification as “Myocardial
Infarction”

Paths are dynamic while being transparent to end user
 Nodes can be added or removed without having an
effect on the user
 Robust among dense networks
We must make network aware of newly added data in an
efficient manner
Martin-14
Discovery Service
CSE
5810

Maps out data on network (similar to DNS)
Martin-15
Mobility with IP Networking
CSE
5810

Illustration through
example:
 Pre-loading
patient’s data
 Large data files
 Based on IP
network
Martin-16
Mobility with IP Networking
CSE
5810


Provider must request
file again
Additional stress to
hospital network
Martin-17
Mobility with Named Data Networking
CSE
5810

Illustration through
example:
 NDN based network
Martin-18
Mobility with Named Data Networking
CSE
5810

Data content cached
Martin-19
Mobility with Named Data Networking
CSE
5810


Data requested again
Reduced redundant data
packets
Martin-20
Simulation Settings
CSE
5810



Ns3 and ndnSIM extension used
Regular IP based network vs. NDN integrated network
Focus:
 Transmission times
 Network stress
Martin-21
Preliminary Simulation Data
CSE
5810
Network Stress
Overall Transmission Time
Martin-22
Conclusion

CSE
5810


Apply NDN concepts in hospital infrastructure
 “What” data instead of “Where”
 Reduce stress on keynote features
 Less bandwidth usage
 Friendlier to mobile devices
Additional features
 Adaptability with discovery service
 Integrated security through data
Challenges
 Acceptability by healthcare
 Ensuring security of data
Martin-23
CSE
5810
THANK YOU
Martin-24