virtual network embeddings with good and bad intentionsstefan/vnet-aalto.pdf · virtual network...
TRANSCRIPT
![Page 1: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/1.jpg)
Virtual Network Embeddings
with Good and Bad Intentions
Stefan Schmid February, 2014
![Page 2: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/2.jpg)
2 Telekom Innovation Laboratories Stefan Schmid
VNets: Virtual Networking Cloud Resources.
![Page 3: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/3.jpg)
3 Telekom Innovation Laboratories Stefan Schmid
Wide-Area VNet: Distributed Cloud.
Service deployment, bandwidth guarantees, ...
![Page 4: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/4.jpg)
4 Telekom Innovation Laboratories Stefan Schmid
Datacenter VNet: Predictable Performance.
Today: only VMs come with performance isolation (if at all).
![Page 5: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/5.jpg)
5 Telekom Innovation Laboratories Stefan Schmid
Datacenter VNet: Predictable Performance.
Today: only VMs come with performance isolation (if at all).
![Page 6: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/6.jpg)
6 Telekom Innovation Laboratories Stefan Schmid
Datacenter VNet: Predictable Performance.
Today: only VMs come with performance isolation (if at all). Without network guarantees: unpredictable, varying, costly application performance.
![Page 7: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/7.jpg)
7 Telekom Innovation Laboratories Stefan Schmid
Virtualization and VNet
Virtualization trend starts to spill-over to the network. Benefit: decouple service from physical constraints, supports flexible embeddings and seamless migrations.
Server virtualization: VMs, OpenStack, ...
Link virtualization: SDN, MPLS, ...
![Page 8: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/8.jpg)
8 Telekom Innovation Laboratories Stefan Schmid
A Graph Embedding Problem!
Physical infrastructure
(e.g., accessed by mobile clients)
Specification:
1. close to mobile clients
2. >100 kbit/s bandwidth for synchronization
Provider 1
Provider 2
CloudNet requests
VNet 2: Mobile service w/ QoS VNet 1: Computation
Specification:
1. > 1 GFLOPS per node
2. Monday 3pm-5pm
3. multi provider ok
![Page 9: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/9.jpg)
9 Telekom Innovation Laboratories Stefan Schmid
Our Prototype Architecture.
Roles in CloudNet Arch.
Service Provider (SP)
(services over the top: knows applications)
Physical Infrastructure Provider (PIP)
(resource, bitpipes: knows demand&infrastructure)
knows service
knows network
![Page 10: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/10.jpg)
10 Telekom Innovation Laboratories Stefan Schmid
Our Prototype Architecture.
Roles in CloudNet Arch.
Service Provider (SP)
(services over the top: knows applications)
Physical Infrastructure Provider (PIP)
(resource, bitpipes: knows demand&infrastructure)
Provide L2 topology: resource and management interfaces, provides indirection layer, across PIPs!
Can be recursive.
Virtual Network Provider (VNP)
(resource broker, compiles resources)
VNet
![Page 11: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/11.jpg)
11 Telekom Innovation Laboratories Stefan Schmid
Our Prototype Architecture.
Roles in CloudNet Arch.
Service Provider (SP)
(services over the top: knows applications)
Physical Infrastructure Provider (PIP)
(resource, bitpipes: knows demand&infrastructure)
Virtual Network Provider (VNP)
(resource broker, compiles resources)
Build upon layer 2: clean slate!
Tailored towards application (OSN, …): routing, addressing, multi-path/redundancy…
E.g., today’s Internet.
Virtual Network Operator (VNO)
(operates VNet, Layer 3+, triggers migration)
Innovation!
![Page 12: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/12.jpg)
12 Telekom Innovation Laboratories Stefan Schmid
Our Prototype Architecture.
API
API
API
AP
Is: e
.g., p
rovis
ion
ing
inte
rfa
ce
s (
mig
ratio
n)
Roles in CloudNet Arch.
Service Provider (SP)
(offers services over the top)
Virtual Network Operator (VNO)
(operates VNet, Layer 3+, triggers migration)
Physical Infrastructure Provider (PIP)
(resource and bit pipe provider)
Virtual Network Provider (VNP)
(resource broker, compiles resources)
Embed!
Embed!
Embed!
![Page 13: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/13.jpg)
13 Telekom Innovation Laboratories Stefan Schmid
Our Prototype Architecture.
API
API
API
AP
Is: e
.g., p
rovis
ion
ing
inte
rfa
ce
s (
mig
ratio
n)
Roles in CloudNet Arch.
Service Provider (SP)
(offers services over the top)
Virtual Network Operator (VNO)
(operates CloudNet, Layer 3+, triggers migration)
Physical Infrastructure Provider (PIP)
(resource and bit pipe provider)
Virtual Network Provider (VNP)
(resource broker, compiles resources)
Embed!
Embed!
Embed!
Internships...
![Page 14: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/14.jpg)
14 Telekom Innovation Laboratories Stefan Schmid
Talk Outline:
1. Which VNets to accept? (And how to embed?) 2. Threats: VNet embeddings with bad intentions? 3. Migrating VNets 4. VNets with in-network processing
![Page 15: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/15.jpg)
15 Telekom Innovation Laboratories Stefan Schmid
Competitive Access Control: Model (1)
Physical Infrastructure
CPU, location, ...
link capacity
Helsinki (location,CPU)
bandwidth
VNet (VPN-like, single resource)
![Page 16: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/16.jpg)
16 Telekom Innovation Laboratories Stefan Schmid
VNet
100 $
Specification of CloudNet request: - terminal locations to be connected - benefit if CloudNet accepted (all-or-nothing, no preemption) - desired bandwidth and allowed traffic patterns - a routing model - duration (from when until when?)
If VNets with these specifications arrive over time, which ones to accept online?
Competitive Access Control: Model (2)
Objective: maximize sum of benefits of accepted VNets
![Page 17: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/17.jpg)
17 Telekom Innovation Laboratories Stefan Schmid
Competitive Access Control: Model (3)
Which ones to accept?
![Page 18: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/18.jpg)
18 Telekom Innovation Laboratories Stefan Schmid
VNet Specifications (1): Traffic Model.
![Page 19: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/19.jpg)
19 Telekom Innovation Laboratories Stefan Schmid
VNet Specifications (2): Routing Model.
![Page 20: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/20.jpg)
20 Telekom Innovation Laboratories Stefan Schmid
VNet Specifications (2): Routing Model.
Relay nodes may add to embedding costs! (resources depend, e.g., on packet rate)
![Page 21: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/21.jpg)
21 Telekom Innovation Laboratories Stefan Schmid
Competitive Embeddings.
Online algorithms make decisions at time t without any knowledge of inputs / requests at times t’>t.
Online Algorithm
Competitive analysis framework:
An r-competitive online algorithm ALG gives a worst-case performance guarantee: the performance is at most a factor r worse than an optimal offline algorithm OPT!
Competitive Analysis
Competitive ratio r, r = Cost(ALG) / cost(OPT) The price of not knowing the future!
Competitive Ratio
No need for complex predictions but still good!
![Page 22: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/22.jpg)
22 Telekom Innovation Laboratories Stefan Schmid
Buchbinder&Naor: Primal-Dual Approach.
Algorithm design and analysis follows online primal-dual approach by Buchbinder&Naor! (Application to general VNet embeddings, traffic&routing models, router loads, duration, approx oracles, ...)
1. Formulate dynamic primal (covering) and dual (packing) LP
2. Derive algorithm which always produces feasible primal solutions and where Primal >= 2*Dual
![Page 23: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/23.jpg)
23 Telekom Innovation Laboratories Stefan Schmid
Result.
Theorem
The presented online algorithm log-competitive
in the amount of resources in the physical network.
(If capacities can be exceeded by a log factor, it is even
constant competitive.)
However, competitive ratio also depends on max benefit.
![Page 24: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/24.jpg)
24 Telekom Innovation Laboratories Stefan Schmid
Algorithm and Proof Sketch (1).
Embedding oracle: algo invokes an oracle procedure to
determine cost of VNet embedding!
![Page 25: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/25.jpg)
25 Telekom Innovation Laboratories Stefan Schmid
Algorithm and Proof Sketch (1).
If resource cost lower than benefit: accept!
![Page 26: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/26.jpg)
26 Telekom Innovation Laboratories Stefan Schmid
Algorithm and Proof Sketch (1).
update allocations for
accepted VNet...
![Page 27: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/27.jpg)
27 Telekom Innovation Laboratories Stefan Schmid
Algorithm and Proof Sketch (1).
otherwise reject
(no change in substrate)
![Page 28: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/28.jpg)
28 Telekom Innovation Laboratories Stefan Schmid
Algorithm and Proof Sketch (1).
Algorithm efficient... except for oracle (static, optimal embedding)!
What if we only use a suboptimal embedding here?
otherwise reject
(no change in substrate)
![Page 29: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/29.jpg)
29 Telekom Innovation Laboratories Stefan Schmid
Algorithm and Proof Sketch (2).
Problem: computation of optimal embeddings NP-hard!
Thus: use approximate embeddings! (E.g., Steiner tree)
GIPO: Embedding approx.:
<insert your favorite
approx algo>
Competitive ratio Approx ratio r
Lemma
The approximation does not reduce the overall competitive
ratio by much: we get *r ratio!
![Page 30: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/30.jpg)
30 Telekom Innovation Laboratories Stefan Schmid
Talk Outline:
1. Which VNets to accept? (And how to embed?) 2. Threats: VNet embeddings with bad intentions? 3. Migrating VNets 4. VNets with in-network processing
![Page 31: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/31.jpg)
31 Telekom Innovation Laboratories Stefan Schmid
Flexible Embeddings: Beyond VPN Model.
Physical Infrastructure
CPU, location, ...
link capacity
resources/ location/...
bandwidth
VNet (general)
![Page 32: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/32.jpg)
32 Telekom Innovation Laboratories Stefan Schmid
Security Issues.
Are VNet embeddings a threat for ISPs?
Do embeddings leak information about infrastructure?
?
![Page 33: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/33.jpg)
33 Telekom Innovation Laboratories Stefan Schmid
Request Complexity.
Are VNet embeddings a threat for ISPs?
How many embeddings needed to fully reveal topology?
Request Complexity
?
?
?
Yes
Yes
No
![Page 34: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/34.jpg)
34 Telekom Innovation Laboratories Stefan Schmid
Embedding Model.
unit capacity
unit capacity
arbitrary node demand arbitrary link
demand
![Page 35: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/35.jpg)
35 Telekom Innovation Laboratories Stefan Schmid
Embedding Model.
unit capacity
unit capacity
arbitrary node demand arbitrary link
demand
relay cost ε>0 (e.g., packet rate)
![Page 36: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/36.jpg)
36 Telekom Innovation Laboratories Stefan Schmid
Embedding Model.
unit capacity
unit capacity
arbitrary node demand arbitrary link
demand
relay cost ε>0 (e.g., packet rate) We will ask for unit capacities on
nodes and links! Essentially a graph immersion problem: disjoint paths for virtual links...
![Page 37: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/37.jpg)
37 Telekom Innovation Laboratories Stefan Schmid
Some Properties Simple...
«Is the network 2-connected?»
?
No
![Page 38: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/38.jpg)
38 Telekom Innovation Laboratories Stefan Schmid
Example: Tree.
How to discover a tree? Graph growing: 1. Test whether triangle fits? (loop-free) 2. Try to add neighbors to node as long as possible, then continue with other node
v
![Page 39: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/39.jpg)
39 Telekom Innovation Laboratories Stefan Schmid
Example: Tree.
How to discover a tree? Graph growing: 1. Test whether triangle fits? (loop-free) 2. Try to add neighbors to node as long as possible, then continue with other node
![Page 40: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/40.jpg)
40 Telekom Innovation Laboratories Stefan Schmid
Example: Tree.
How to discover a tree? Graph growing: 1. Test whether triangle fits? (loop-free) 2. Try to add neighbors to node as long as possible, then continue with other node
![Page 41: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/41.jpg)
41 Telekom Innovation Laboratories Stefan Schmid
Example: Tree.
How to discover a tree? Graph growing: 1. Test whether triangle fits? (loop-free) 2. Try to add neighbors to node as long as possible, then continue with other node
![Page 42: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/42.jpg)
42 Telekom Innovation Laboratories Stefan Schmid
Example: Tree.
How to discover a tree? Graph growing: 1. Test whether triangle fits? (loop-free) 2. Try to add neighbors to node as long as possible, then continue with other node
![Page 43: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/43.jpg)
43 Telekom Innovation Laboratories Stefan Schmid
Example: Tree.
How to discover a tree? Graph growing: 1. Test whether triangle fits? (loop-free) 2. Try to add neighbors to node as long as possible, then continue with other node
Virtual links may be embedded over multiple physical links!
![Page 44: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/44.jpg)
44 Telekom Innovation Laboratories Stefan Schmid
Tree Solution: Graph Growing.
TREE ALGORITHM: line strategy 1. Binary search on longest path («anchor»):
...
2. Last and first node explored, explore «branches» at pending nodes
Amortized Analysis: Per discovered physical link at most one query, plus at most one per physical node (no incident links).
![Page 45: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/45.jpg)
45 Telekom Innovation Laboratories Stefan Schmid
Greedy Graph Growing on General Graphs? (1)
Finding path...
![Page 46: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/46.jpg)
46 Telekom Innovation Laboratories Stefan Schmid
Greedy Graph Growing on General Graphs? (1)
Finding neighbors...
![Page 47: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/47.jpg)
47 Telekom Innovation Laboratories Stefan Schmid
Greedy Graph Growing on General Graphs? (1)
Finding more neighbors...
![Page 48: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/48.jpg)
48 Telekom Innovation Laboratories Stefan Schmid
Greedy Graph Growing on General Graphs? (1)
How to close the gap? Adding connections between existing CloudNet nodes is expensive: try all pairs!
![Page 49: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/49.jpg)
49 Telekom Innovation Laboratories Stefan Schmid
Greedy Graph Growing on General Graphs? (1)
How to close the gap? Adding connections between existing CloudNet nodes is expensive: try all pairs!
![Page 50: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/50.jpg)
50 Telekom Innovation Laboratories Stefan Schmid
Greedy Graph Growing on General Graphs? (2)
Simple solution: First try to find the «knitting»! - The «two-or-more» connected components - Later «expand nodes» and «expand edges»
![Page 51: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/51.jpg)
51 Telekom Innovation Laboratories Stefan Schmid
Simple solution: First try to find the «knitting»! - The «two-or-more» connected components - Later «expand nodes» and «expand edges»
Greedy Graph Growing on General Graphs? (2)
![Page 52: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/52.jpg)
52 Telekom Innovation Laboratories Stefan Schmid
Greedy Graph Growing on General Graphs? (3)
Idea: Ask graph «motif» only if it’s guaranteed that it cannot be embedded over a more highly connected subgraph! (And connectivity has to be added later.)
1 2 3 4 5
6 7
8 2
1 3 4 5
6 7
8
Relay cost: 4 ε
A B BB
Careful: What goes first also depends on entire motif sequences!
A B B A A BB
![Page 53: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/53.jpg)
53 Telekom Innovation Laboratories Stefan Schmid
Remark.
Minor vs embedding:
Even with unit link capacity, for small epsilon, graph A may be embeddable ( ) into graph B although A is not a minor of B!
Planar graph (and hence K5-minor free): But K5 can be embedded here!
Graph A is a minor of B if A can be obtained from B by (1) deleting nodes, (2) deleting edges, or (3) contracting two nodes along edges.
Graph Minor A
B
![Page 54: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/54.jpg)
54 Telekom Innovation Laboratories Stefan Schmid
Dictionary Attack: Expansion Framework.
Examples Tree motifs: Cactus motifs:
Basic “knittings” of the graph.
Motif Define an order on motif sequences: Constraints on which sequence to ask first in order not to overlook a part of the topology. (E.g., by embedding links across multiple hops.)
Dictionary
Poset
Framework
Explore branches according
to dictionary order,
exploiting poset property.
Poset = partially ordered set (1) Reflexive: G G (2) Transitive: G G’ and G’ G’’, then G G’’ (3) Antisymmetric: G G’ and G’ G implies G=G’ (isomorphic)
![Page 55: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/55.jpg)
55 Telekom Innovation Laboratories Stefan Schmid
query order
Dictionary Attack: Expansion Framework.
solves, e.g.,
Dictionary dag (for chain C, cycle Y, diamond D, ...) with attachment points:
poset poset
Complexity: Depends on dictionary depth and number of attachment points
![Page 56: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/56.jpg)
56 Telekom Innovation Laboratories Stefan Schmid
Overview of Results.
PIP 3 Tree Can be explored in O(n)
requests. This is optimal!
PIP 3 General Graph Can be explored in O(n2)
requests. This is optimal!
Cactus Graph Can be explored in O(n)
requests. This is optimal!
Lower bound: via number of possible trees and binary information.
Idea: Make spanning tree and then try all edges. (Edges directly does not work!)
Via «graph motifs»! A general framework exploiting poset relation.
![Page 57: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/57.jpg)
57 Telekom Innovation Laboratories Stefan Schmid
Overview of Results.
PIP 3 Tree Can be explored in O(n)
requests. This is optimal!
PIP 3 General Graph Can be explored in O(n2)
requests. This is optimal!
Cactus Graph Can be explored in O(n)
requests. This is optimal!
Lower bound: via number of possible trees and binary information.
Idea: Make spanning tree and then try all edges. (Edges directly does not work!)
Via «graph motifs»! A general framework exploiting poset relation.
![Page 58: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/58.jpg)
58 Telekom Innovation Laboratories Stefan Schmid
Dictionary Attacks: Expand Framework.
Examples Tree motifs: Cactus motifs:
Basic “knittings” of the graph.
Motif Define an order on motif sequences: Constraints on which sequence to ask first in order not to overlook a part of the topology. (E.g., by embedding links across multiple hops.)
Dictionary
Partially ordered set: embedding relation fulfills reflexivity, anti-symmetry, transitivity.
Poset
Framework
Explore branches according
to dictionary order,
exploiting poset property.
![Page 59: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/59.jpg)
59 Telekom Innovation Laboratories Stefan Schmid
Talk Outline:
1. Which VNets to accept? (And how to embed?) 2. Threats: VNet embeddings with bad intentions? 3. Migrating VNets 4. VNets with in-network processing
![Page 60: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/60.jpg)
60 Telekom Innovation Laboratories Stefan Schmid
The Virtual Service Migration Problem.
on service!
const bw
Given a virtual network with guaranteed bandwidth: where to migrate service?
![Page 61: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/61.jpg)
61 Telekom Innovation Laboratories Stefan Schmid
The Virtual Service Migration Problem.
Given a virtual network with guaranteed bandwidth: where to migrate service?
Simple model: one service, constant migration cost (interruption), access along graph.
Cost: m * # migrations + sum of access latency.
on service!
const bw
![Page 62: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/62.jpg)
62 Telekom Innovation Laboratories Stefan Schmid
The Virtual Service Migration Problem.
on service!
const bw
Interesting variant of Metrical Task System:
Migration cost depends on available bandwidth, while access cost depends on graph distance. If migration cost constant: uniform metrical task system with access on graph (triangle inequality).
Given a virtual network with guaranteed bandwidth: where to migrate service?
Simple model: one service, constant migration cost (interruption), access along graph.
Cost: m * # migrations + sum of access latency.
![Page 63: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/63.jpg)
63 Telekom Innovation Laboratories Stefan Schmid
Algorithms?
![Page 64: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/64.jpg)
64 Telekom Innovation Laboratories Stefan Schmid
Randomized Algo:
1. Access cost counters at each node (if service there)
2. When counter exceeds m, migrate to random node with
counter lower than m.
3. When no node left, epoch ends. Reset and restart.
on service!
COUNT(v)
COUNT(v)
COUNT(v)
COUNT(v)
COUNT(v)
COUNT(v)
![Page 65: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/65.jpg)
65 Telekom Innovation Laboratories Stefan Schmid
Randomized Algo:
1. Access cost counters at each node (if service there)
2. When counter exceeds m, migrate to random node with
counter lower than m.
3. When no node left, epoch ends. Reset and restart.
on service!
COUNT(v)
COUNT(v)
COUNT(v)
COUNT(v)
COUNT(v)
COUNT(v) Analysis: log(n)-competitive
Offline cost per epoch:
at least m (migrate or access cost)
Online cost per epoch:
Per phase at most 2m (access plus migration).
At most log(n) phases: go to random node in remaining order.
![Page 66: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/66.jpg)
66 Telekom Innovation Laboratories Stefan Schmid
Deterministic Algo:
1. Access cost counters at each node (if service there)
2. When counter exceeds m, deactivate nodes with counters > m/40, migrate to active node in center of active component.
3. When no node left, epoch ends. Reset and restart.
on service!
COUNT(v)
COUNT(v)
COUNT(v)
COUNT(v)
COUNT(v)
COUNT(v) Analysis: log(n)-
competitive
Offline cost per epoch:
at least m (migrate or access cost)
Online cost per epoch:
Per phase at most 2m (access plus migration).
At most log(n) phases: exploit triangle inequality.
![Page 67: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/67.jpg)
67 Telekom Innovation Laboratories Stefan Schmid
Center-of-Gravity Algo: Example.
67
Before phase 1:
active
inactive
on service!
![Page 68: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/68.jpg)
68 Telekom Innovation Laboratories Stefan Schmid
Center-of-Gravity Algo: Example.
68
Before phase 2:
active
inactive
on service!
![Page 69: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/69.jpg)
69 Telekom Innovation Laboratories Stefan Schmid
Center-of-Gravity Algo: Example.
69
End of epoch:
active
inactive
on service!
![Page 70: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/70.jpg)
70 Telekom Innovation Laboratories Stefan Schmid
Center-of-Gravity Algo: Result.
Competitive analysis? Assume constant bandwidths!
r = ALG / OPT ?
Lower bound cost of OPT:
In an epoch, each node has
at least access cost m, or
there was a migration of cost m.
Upper bound cost of ALG:
We can show that each phase
has cost at most 2m (access
plus migration), and there are
at most log(n) many phases
per epoch!
Theorem
ALG is log(n) competitive! A special uniform metrical task system (graph metric for access)!
![Page 71: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/71.jpg)
71 Telekom Innovation Laboratories Stefan Schmid
Optimality?
Theorem «Center of Gravity» algorithm is
log(n) competitive!
Also a much simpler randomized
algorithm achieves this!
log(n)/loglog(n) lower bound follows from online function tracking reduction!
on service!
F(x)
Online function tracking with linear penalties: Alice observes values xt and Bob has representative value F(x). Upon new value, either Alice transmits (migration cost) or pays difference |xt-y| (access cost).
![Page 72: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/72.jpg)
72 Telekom Innovation Laboratories Stefan Schmid
Optimality?
Theorem «Center of Gravity» algorithm is
log(n) competitive!
Also a much simpler randomized
algorithm achieves this!
log(n)/loglog(n) lower bound follows from online function tracking reduction!
on service!
F(x)
Can be achieved with a refined analysis!
![Page 73: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/73.jpg)
73 Telekom Innovation Laboratories Stefan Schmid
The Online Algorithm FOLLOWER.
Simplified Follower
1. 1. Fi are requests handled while service at fi
2. 2. to compute fi+1 (new pos), Follower only takes into account requests during fi: Fi
3. 3. migrate to center of gravity of Fi, as soon as migration costs there are amortized (and «reset counters» immediately)!
Concepts: - Learn from the past: migrate to center of gravity of best
location in the past
- Amortize: migrate only when access cost at current
node is as high as migration cost!
![Page 74: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/74.jpg)
74 Telekom Innovation Laboratories Stefan Schmid
The Online Algorithm FOLLOWER.
Concepts: - Learn from the past: migrate to center of gravity of best
location in the past
- Amortize: migrate only when access cost at current
node is as high as migration cost!
Simplified Follower
1. 1. Fi are requests handled while service at fi
2. 2. to compute fi+1 (new pos), Follower only takes into account requests during fi: Fi
3. 3. migrate to center of gravity of Fi, as soon as migration costs there are amortized (and «reset counters» immediately)!
![Page 75: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/75.jpg)
75 Telekom Innovation Laboratories Stefan Schmid
Intuition.
75
on service!
![Page 76: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/76.jpg)
76 Telekom Innovation Laboratories Stefan Schmid
Intuition.
76
on service! Fi
= fi
![Page 77: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/77.jpg)
77 Telekom Innovation Laboratories Stefan Schmid
Intuition.
77
on service!
Fi
= fi+1
![Page 78: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/78.jpg)
78 Telekom Innovation Laboratories Stefan Schmid
Modeling Access and Migration Costs.
Access Costs 1. Latency along shortest path in graph.
2. (Graph distances, and in particular: metric!)
Migration Costs 1. Generalized models:
- E.g., depends on bandwidth along path (duration of service interruption)
- E.g., depends on distance travelled
(latency)
- Discount: e.g., VNP (number of
migrations, distance travelled, ...)
![Page 79: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/79.jpg)
79 Telekom Innovation Laboratories Stefan Schmid
Modeling Access and Migration Costs.
Access Costs 1. Latency along shortest path in graph.
2. (Graph distances, and in particular: metric!)
Migration Costs 1. Generalized models:
- E.g., depends on bandwidth along path (duration of service interruption)
- E.g., depends on distance travelled
(latency)
- Discount: e.g., VNP (number of
migrations, distance travelled, ...)
![Page 80: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/80.jpg)
80 Telekom Innovation Laboratories Stefan Schmid
Competitive Ratio of FOLLOWER.
Competitive analysis? FOLLOWER / OPT?
Theorem If no discounts are given,
Follower is log(n)/loglog(n) competitive!
Theorem If migration costs depend on travelled
distance (page migration), competitive
ratio is O(1), even with discounts.
Simple model with migration
costs = bandwidth, and
homogeneous
Page migration model with
migration costs = distance,
but discounts
![Page 81: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/81.jpg)
81 Telekom Innovation Laboratories Stefan Schmid
Related Work.
- Metrical Task Systems:
- Classical online problem where server at certain location («state») serves requests at certain costs; state transitions also come at certain costs («migration»)
- Depending on migration cost function more general (we have graph access costs) and less general (we allow for migration discounts)
- E.g., uniform space metrical task system: migration costs constant, but access costs more general than graph distances! Lower bound of log(n) vs log(n)/loglog(n) upper bound in our case.
- Online Page Migration
- Classical online problem from the 80ies; we generalize cost function to distance discounts, while keeping O(1)-competitive
Our work lies between!
![Page 82: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/82.jpg)
82 Telekom Innovation Laboratories Stefan Schmid
Talk Outline:
1. Which VNets to accept? (And how to embed?) 2. Threats: VNet embeddings with bad intentions? 3. Migrating VNets 4. VNets with in-network processing
![Page 83: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/83.jpg)
83 Telekom Innovation Laboratories Stefan Schmid
VNet with Processing
Unicast: one connection to each receiver (same for aggregation)!
![Page 84: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/84.jpg)
84 Telekom Innovation Laboratories Stefan Schmid
VNet with Processing
Multicast: processing / splitting at each node
![Page 85: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/85.jpg)
85 Telekom Innovation Laboratories Stefan Schmid
VirtuCast: New Tradeoff?
Multicast Unicast
![Page 86: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/86.jpg)
86 Telekom Innovation Laboratories Stefan Schmid
VirtuCast: Optimal Tradeoff
![Page 87: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/87.jpg)
87 Telekom Innovation Laboratories Stefan Schmid
A Single-Commodity Algorithm
![Page 88: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/88.jpg)
88 Telekom Innovation Laboratories Stefan Schmid
A Single-Commodity Algorithm
VirtuCast: 2 Stages
1. Compute single-commodity MIP
2. Make flow decomposition: find path
![Page 89: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/89.jpg)
89 Telekom Innovation Laboratories Stefan Schmid
VNets. Use Cases
Datacenters „VPN++“
Migration / Service Deployment Spillover/Out-Sourcing Goal: Move with the sun, with the commuters, (QoS) allow for maintenance, avoid roaming costs…: e.g., SAP/game/translator server, small CDN server…
Goal: Fully specified CloudNet mapping constraints (e.g., end-points for a telco), but with QoS guarantees (e.g., bandwidth) along links
Berlin
Tel Aviv Palo Alto
1Mbit/s 1Mbit/s
1Mbit/s
„November 22,
1pm-2pm!“
Berlin
„any European cloud provider (e.g. due to legal issues?)“
Berlin (corporate access network)
< 50ms
< 50ms
„50 TB storage, 10 Tflops computation!“
any
any any < 10ms
> 100 MB/s
„Guaranteed resources, job deadlines met, no overhead!“
“Network may delay execution: costly for per hour priced VM!”
< 10ms
> 100 MB/s < 10ms > 100 MB/s
Elastic computing
See, e.g., Octopus system (SIGCOMM 2011) Thank you!
![Page 90: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/90.jpg)
90 Telekom Innovation Laboratories Stefan Schmid
Backup.
![Page 91: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/91.jpg)
91 Telekom Innovation Laboratories Stefan Schmid
Proof Sketch (1): Simplified LP.
maximize
benefit!
realization of i-th
request (will be integer,
accept fully or not at all)
... while ensuring
capacity and
no more than demand!
![Page 92: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/92.jpg)
92 Telekom Innovation Laboratories Stefan Schmid
Proof Sketch (2): Simplified LP. essentially, exponential load...
![Page 93: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/93.jpg)
93 Telekom Innovation Laboratories Stefan Schmid
Proof Sketch (3): Simplified LP.
oracle
(triangle only)
update primal
variables if accepted
![Page 94: Virtual Network Embeddings with Good and Bad Intentionsstefan/vnet-aalto.pdf · Virtual Network Provider (VNP) (resource broker, compiles resources) Build upon layer 2: clean slate!](https://reader033.vdocuments.net/reader033/viewer/2022060605/605ab984bc99c5334127ec2a/html5/thumbnails/94.jpg)
94 Telekom Innovation Laboratories Stefan Schmid
Proof Sketch (4): Simplified LP.
after each request,
primal variables
constitute feasible
solutions...