Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

22
Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu

Transcript of Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

Page 1: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

Drafting Behind Akamai (Travelocity-Based Detouring)

Dr. Yingwu Zhu

Page 2: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

Motivation

• Overlay networks– Build the “view” of the underlying network– Passive and active measurements• To determine overlay paths according to some metric,

e.g., latency?• Redundant, non-scalable, & overhead (expensive!)

• What common service we may need in place?– An Internet weather-report service

Page 3: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

Proposed Approach• Fact– CDNs (e.g., Akamai) perform extensive network and

server measurements• Publish the results through DNS over short time scales

• Can overlay networks recycle measurements collected by production CDNs?– Significantly reduce the amount of measurements (a

complementary service) – No new infrastructure need to be deployed– Inherit the robustness of DNS– Easy integration with existing systems

Page 4: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

CDN-Driven One-Hop Source Routing

S

A1

An

A2

……

..

D

DNS Server

En

E2

E1

Page 5: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

Key Questions

• How does Akamai work?– DNS translation– How many web replicas does a client see?– Impact of different sites (e.g., Yahoo vs. NY Times)?– DNS redirection dynamics?– Network or server latency?

• An example application: one-hop source routing– Mapping CDN servers to overlay nodes– Low-overhead protocols for exploiting CDN

redirections

Page 6: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

6

DNS “Black Magic”

WebClient

Akamai High-LevelDNS Server

Akamai Low-LevelDNS Server

Local DNS ServerPCWorld AuthoritativeDNS Server

PCWorld Web Server

AkamaiEdge Server

Root DNS Server

Top-Level Domain DNS Server

www.pcworld.com

www.pcworld.comwww.pcworld.com

www.pcworld.com

http request/response

images.pcworld.com

CNAME: a1694.g.akamai.netimages.pcworld.com

a1694.g.akmai.net

a1694.g.akmai.net2 ip addresses of Akamai Edge Servers

2 ip addresses of Akamai Edge Servers

fetch image files

Page 7: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

7

Measuring Akamai

• 2-months long measurement• 140 PlanetLab (PL) nodes

– 50 US and Canada, 35 Europe, 18 Asia, 8 South America, the rest randomly scattered

• Every 20 sec, each PL node queries an appropriate CNAME for– Yahoo, CNN, Fox News, NY Times, etc.

PL Node

AkamaiEdge Server 2

AkamaiEdge Server 3

Akamai Low-LevelDNS Server

.……

AkamaiEdge Server 1

Page 8: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

8

Initial Results

day

night

Berkeley Purdue

Page 9: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

9

Server Diversity for Yahoo

Good overlay-to-CDN mapping candidates

Majority of PL nodessee between 10 and 50 Akamai edge-servers

Nodes far away from Akamaihot-spots

Page 10: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

10

Multiple Akamai Customers

Amazon is not hosted on UMass and LBNL

Yahoo and NYTimes are not hosted in U. Oregon

FEMA is poor!

Page 11: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

11

Redirection Dynamics

Berkeley Brazil

Korea

Page 12: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

12

Key Questions

• How does Akamai work?– DNS translation– How many web replicas does a client see?– Impact of different sites (e.g., Yahoo vs. NY Times)?– DNS redirection dynamics?– Network or server latency?

• Potentials for one-hop source routing?– Mapping CDN servers to overlay nodes– Low-overhead protocols for exploiting CDN

redirections

Page 13: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

13

Methodology

Planet Lab Node

Akamai Low-LevelDNS Server

10 Best Akamai Edge Servers

………

Page 14: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

14

Do CDN redirections correlate to network latencies? (1)

• Rank = r1+r2-1– 16 means perfect correlation

– 0 means poor correlationBrazil is poor

MIT and Amsterdam are excellent

Page 15: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

15

Do CDN redirections correlate to network latencies? (2)

CNN is poor!

70% paths are among best 10%

97% better than average

Page 16: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

16

Akamai-Driven One-Hop Source Routing

S

A1

An

A2

……

..

D

DNS Server

Redirection dynamics sufficiently small for network control

En

E2

E1

Redirections driven by network conditions

Potential for CDN-to-overlay mapping

Page 17: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

17

Methodology

Source

Akamai Low-LevelDNS Server

Destination

Direct PathAkamai path 3

Akamai path 1

……

....

.

Akamai path 10

Akamai path 2

Akamai path 1

Akamai path 2

Akamai path 3

Akamai path 10

Page 18: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

18

Akamai-driven source routing (1)

80% Taiwan15% Japan5 % U.S.

75% U.K.25% U.S.

Taiwan-UK UK-Taiwan

Page 19: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

19

Akamai-driven source routing (2)

Direct paths betterthan Akamai paths

25% of Akamaipaths better thandirect pathsAkamai and direct

paths have similar performance

Experiment: US (6), Europe (3), S. America (2), Asia (3)

Page 20: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

20

Path pruning

• Fact– Not always is Akamai-driven path better than the

direct one

• Practical issues– How frequently to make a decision whether to use

the direct or the Akamai path?– Should one use • the first (of the 2 paths) returned by Akamai (FAS)• the better (of the 2 paths) returned by Akamai (BTAS)

Page 21: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

21

Path Pruning Result

Direct path accounts for 78% of the gain

2 hours update frequency before the performance declines

BTAS better than FAS

Page 22: Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.

22

Conclusions

• Reuse measurements performed by CDNs• Reverse-engineering Akamai– DNS redirections sufficiently small– Strong correlation to network conditions– All clients see a large number of paths

• CDN-driven one-hop source routing– 25% of Akamai paths outperform direct paths– 50% of nodes “discovered” by Akamai outperform

direct paths– Low-overhead pruning algorithms

• Global Internet “weather-report” service for little to no cost