Download Identity Management and Resource Allocation in the Network

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Distributed firewall wikipedia , lookup

Cracking of wireless networks wikipedia , lookup

IEEE 802.1aq wikipedia , lookup

Zero-configuration networking wikipedia , lookup

Piggybacking (Internet access) wikipedia , lookup

CAN bus wikipedia , lookup

Computer network wikipedia , lookup

Asynchronous Transfer Mode wikipedia , lookup

Network tap wikipedia , lookup

Recursive InterNetwork Architecture (RINA) wikipedia , lookup

Airborne Networking wikipedia , lookup

Peer-to-peer wikipedia , lookup

Transcript
Identity Management and Resource Allocation
in the Network Virtualization Environment
Mosharaf Chowdhury
School of Computer Science
University of Waterloo
January 21, 2009
1
NETWORK VIRTUALIZATION
January 21, 2009
2
Why Network Virtualization?
• Internet is almost ossified
– Lots of band-aids and makeshift solutions (e.g., overlays)
– A new architecture (aka clean-slate) is needed
• Hard to come up with a one-size-fits-all architecture
– Almost impossible to predict what future might unleash
• Why not create an all-sizes-fit-into-one architecture
instead!
– Open and expandable
– Coexistence of heterogeneous architectures
January 21, 2009
3
What is Network Virtualization?
• Transparent abstraction of networking platform and
resources
– Multiple logical interpretations of the physical characteristics
• Multiple virtual networks (VNs)
• Additional level of indirection
– Indirect access to network resources
• Resource partitioning and isolation
– Physical and logical
– Dynamic provisioning and configuration
January 21, 2009
4
Network Virtualization Environment
January 21, 2009
5
Challenges
Virtual Network
Embedding
Instantiation
– Concerned with issues related to successful
creation of virtual networks
– Deals with operations of virtual networks and
virtual components
Management
– Manages co-existing virtual networks
January 21, 2009
Identity
Management
Operations
6
Identity Management in the Network Virtualization Environment
IMARK
January 21, 2009
7
Motivation
• High level of dynamism
– Macro Level: Merge/Separate VNs
– Micro Level: Add/Join/Migration of end hosts and virtual routers
• Mobility
– Geographical
– Logical
• Überhoming
– Simultaneously connect to multiple InPs and VNs
January 21, 2009
8
Design Principles
1. Separation of Identity and Location
– Inherent support for mobility and Überhoming
2. Local Autonomy
– Flexibility of naming and addressing in different VNs
– Defined interfaces and mechanisms for cooperation
3. Global Identifier Space
– Local identifiers have no end-to-end significance
January 21, 2009
9
iMark Overview
• Concepts
1. Identifier Spaces
2. Mappings
• Components
1. Controllers
2. Adapters
January 21, 2009
10
Operations: Macro Level
• Federation
– Multiple VNs create
common administrative
domain
– Controller network
• Hierarchy
– Aggregation of mappings in
representative controllers
– Balanced and unbalanced
January 21, 2009
11
Operations: Micro Level
• Join
– Add mappings
• Lookup and Connection Setup
– State setup in the network
• Leave
– Remove mappings
• Mobility
– Soft handoff
January 21, 2009
12
Evaluation
Mean Mapping Size Per Controller
January 21, 2009
Mean Lookups Resolved
13
Intra-domain Resource Allocation through Virtual Network Embedding
VINEYARD
January 21, 2009
14
Virtual Network Embedding
10
80
a
10
12
b
c
10
10
A
55
B
22
12
90
10
C
60
20
15
D
20
15
E
January 21, 2009
5
d
f
20
20
F
85
17
25
e
5
50
10
G
70
17
H
65
15
Substrate Graph Augmentation
a
∞
80
55
15
A
∞
10
a
10
b
10
10
B
22
C
c
60
10
∞
∞
12
90
12
c
∞
D
15
50
10
20
E
F
85
17
25
G
H
17
70
65
∞
∞
b
January 21, 2009
16
D-ViNE and R-ViNE
For each VN request:
– Augment the substrate graph
– Solve the resulting LP
– For each virtual node:
• Calculate the probability for each meta-node to be selected
for the corresponding virtual node
• Selection:
INITIALIZATION
NODE MAPPING
– D-ViNE: Select the meta-node with the highest probability
– R-ViNE: Select a meta-node randomly with the calculated probability
– Use MCF to map virtual edges
LINK MAPPING
– If the VN request is accepted
FINALIZATION
•
January 21, 2009
Update residual capacities of the substrate resources
17
Acceptance Ratio
January 21, 2009
18
Revenue Vs Cost
Revenue
January 21, 2009
Cost
19
Resource Utilization
Node Utilization
January 21, 2009
Link Utilization
20
What have we done? What will we do?
SUMMARY
January 21, 2009
21
Contributions
• Survey of Network Virtualization (Submitted + TechReport)
– Historical perspective
– Categorization of existing projects
– Enumeration of open problems
• Identity Management Framework (IM’09)
– Interoperability with flexibility to support mobility and Überhoming
• Virtual Network Embedding Algorithms (INFOCOM’09)
– Better embedding quality
– Mathematical foundation
January 21, 2009
22
Future Work
• iMark Prototype Development
– Further evaluation
• Theoretical Analysis of D-ViNE and R-ViNE
– Approximation factors
– Economic models
• Inter-domain VN embedding
January 21, 2009
23
Collaborators
• Fida-E Zaheer (iMark)
• Muntasir Raihan Rahman (ViNEYard)
• Network Virtualization Project Members
January 21, 2009
24
Questions?
Mosharaf Chowdhury
http:// www.mosharaf.com/
January 21, 2009
25
BACKUP SLIDES
January 21, 2009
26
Related Concepts
1.
2.
3.
4.
Virtual Local Area Networks (VLAN)
Virtual Private Networks (VPN)
Active and Programmable Networks
Overlay Networks
January 21, 2009
27
Downsides of Overlay Networks
• Largely used as narrow fixes for specific problems
– No holistic view
• Most overlays are designed in the application
layer
– Cannot support radically different concepts
Anderson et al.
January 21, 2009
28
What is a Virtual Network (VN)?
• A collection of virtual nodes and virtual links forming a
virtual topology
– Subset of physical topology
– Basic entity of the NVE
• A virtual node is hosted on a particular physical node
– Multiple virtual nodes can coexist
• A virtual link spans over a physical path
– Includes a portion of the underlying physical resources
January 21, 2009
29
Business Model
Players
•
Infrastructure Providers (InP)
–
•
End User
Create and manage virtual networks
Deploy customized end-to-end services
End Users
–
•
Manage underlying physical networks
Service Providers (SP)
–
–
•
Relationships
Buy and use services from different service
providers
Brokers
–
Mediators/Arbiters
January 21, 2009
SLA
Broker
EIA
Service Provider
NPA
SIA
Infrastructure
Provider
IIA
30
Hierarchy of Roles
January 21, 2009
31
Basic Concepts
Principles
Design Goals
•
•
•
•
•
•
•
•
•
•
•
•
•
Concurrence
Recursion
Inheritance
Revisitation
January 21, 2009
Flexibility
Manageability
Scalability
Isolation
Stability and Convergence
Programmability
Heterogeneity
Experimental and Deployment Facility
Legacy Support
32
What is Network Virtualization? (Revisited)
Network virtualization is a networking environment that allows
multiple service providers to dynamically compose multiple
heterogeneous virtual networks that coexist together in
isolation from each other, and to deploy customized end-toend services on-the-fly as well as manage them on those
virtual networks for the end-users by effectively sharing and
utilizing underlying network resources leased from multiple
infrastructure providers.
January 21, 2009
33
Classification
• Networking technology
– Targeted technology for virtualization
• Layer of virtualization
– Particular layer in the network stack where virtualization is introduced
• Architectural domain
– Specific problem domain that virtualization addresses
• Level of virtualization
– Granularity at which virtualization is realized
January 21, 2009
34
Existing Projects
Project
Architectural Domain
Networking
Technology
VNRMS
Virtual network
management
ATM/IP
Tempest
Enabling alternate
control architectures
ATM
Link
NetScript
Dynamic composition
of services
IP
Network
Node
Genesis
Spawning virtual
network architectures
Network
Node/Link
January 21, 2009
Layer of
Virtualization
Level of
Virtualization
Node/Link
35
Existing Projects (Cont.)
Project
Architectural Domain
VNET
Virtual machine Grid
computing
VIOLIN
Deploying on-demand
value-added services on
IP overlays
X-Bone
Layer of
Virtualization
Level of
Virtualization
Link
Node
IP
Application
Node
Automating deployment
of IP overlays
IP
Application
Node/Link
PlanetLab
Deploy and manage
overlay-based testbeds
IP
Application
Node
UCLP
Dynamic provisioning
and reconfiguration of
lightpaths
SONET
Physical
Link
January 21, 2009
Networking
Technology
36
Existing Projects (Cont.)
Project
Architectural Domain
Networking
Technology
Layer of
Virtualization
AGAVE
End-to-end QoS-aware
service provisioning
IP
Network
GENI
Creating customized
virtual network testbeds
Heterogeneous
VINI
Evaluating protocols
and services in a
realistic environment
CABO
Deploying value-added
end-to-end services on
shared infrastructure
January 21, 2009
Level of
Virtualization
Link
Heterogeneous
Full
37
Major Ongoing Projects
Project
Originated In
Link
4WARD
Europe
http://www.4ward-project.eu/
AKARI
Japan
http://akari-project.nict.go.jp/
CABO
USA
http://www.cs.princeton.edu/~jrex/virtual.html
Clean Slate
USA
http://cleanslate.stanford.edu/
GENI
USA
http://www.geni.net/
NouVeau
Canada
http://netlab.cs.uwaterloo.ca/virtual/
PlanetLab
USA
http://www.planet-lab.org/
Trilogy
Europe
http://www.trilogy-project.org/
UCLP
Canada
http://www.uclp.ca/
VINI
USA
http://www.vini-veritas.net/
January 21, 2009
38
Entities and Identifier Spaces
Entities
1. Service Provider
2. Virtual Network
3. Virtual Resource
4. Infrastructure Provider /
Physical Network
5. Physical Resource
6. End User
January 21, 2009
Identifier Spaces
1. IDS_ISP
2. IDS_VN
3. IDS_VR
4. IDS_PR
5. IDS_EH
39
Relationships between Entities
January 21, 2009
40
Mappings between Different Identifiers
January 21, 2009
41
Sequence Diagram: Join
January 21, 2009
42
Sequence Diagram: Lookup
January 21, 2009
43
January 21, 2009
44
January 21, 2009
45
D-ViNE
January 21, 2009
46
R-ViNE
January 21, 2009
47
Summary of Compared Algorithms
January 21, 2009
48
January 21, 2009
49
January 21, 2009
50
January 21, 2009
51