Anycast Latency How many sites are enough?

Similar documents
Evolution Strategies for Optimizing Rectangular Cartograms

OH BOY! Story. N a r r a t iv e a n d o bj e c t s th ea t e r Fo r a l l a g e s, fr o m th e a ge of 9

The State of the (Romanian) Internet

A L A BA M A L A W R E V IE W

Last 4 Digits of USC ID:

IEC SYSTEM FOR CONFORMITY TESTING AND CERTIFICATION OF ELECTRICAL EQUIPMENT

5. Atmospheric Supply of Mercury to the Baltic Sea in 2015

Nucleus. Electron Cloud

COMMISSION (2001/793/EC) Having regard to the Treaty establishing the European Community,

02/05/09 Last 4 Digits of USC ID: Dr. Jessica Parr

Advanced Placement. Chemistry. Integrated Rates

Parts Manual. EPIC II Critical Care Bed REF 2031

Use precise language and domain-specific vocabulary to inform about or explain the topic. CCSS.ELA-LITERACY.WHST D

Software Process Models there are many process model s in th e li t e ra t u re, s om e a r e prescriptions and some are descriptions you need to mode

Atoms and the Periodic Table

The Periodic Table of the Elements

Forecasting with Bayesian Global Vector Autoregressive Models

CHEM 10113, Quiz 5 October 26, 2011

CHEM 108 (Spring-2008) Exam. 3 (105 pts)

o Alphabet Recitation

POLYTECHNIC OF NAMIBIA

8. Relax and do well.

Element Cube Project (x2)

HANDOUT SET GENERAL CHEMISTRY II

9/20/2017. Elements are Pure Substances that cannot be broken down into simpler substances by chemical change (contain Only One Type of Atom)

(C) Pavel Sedach and Prep101 1

T h e C S E T I P r o j e c t

Radiometric Dating (tap anywhere)

What is the periodic table?

Made the FIRST periodic table

MANY ELECTRON ATOMS Chapter 15

Guide to the Extended Step-Pyramid Periodic Table

GRIPs. Overview Andrea Ćirlićová Business Area Manager, System Development

1 Genesis 1:1. Chapter 10 Matter. Lesson. Genesis 1:1 In the beginning God created the heavens and the earth. (NKJV)

Chemistry 185 Exam #2 - A November 5, Lab Day and Time: Instructions. 1. Do not open the exam until you are told to start.

Ch. 9 NOTES ~ Chemical Bonding NOTE: Vocabulary terms are in boldface and underlined. Supporting details are in italics.

CHEM 107 (Spring-2005) Exam 3 (100 pts)

Chapter 12 The Atom & Periodic Table- part 2

Country

Use precise language and domain-specific vocabulary to inform about or explain the topic. CCSS.ELA-LITERACY.WHST D

7. Relax and do well.

PHYSICAL SCIENCES GRADE : 10

Instructions. 1. Do not open the exam until you are told to start.

The Periodic Table. Periodic Properties. Can you explain this graph? Valence Electrons. Valence Electrons. Paramagnetism

CHEM Come to the PASS workshop with your mock exam complete. During the workshop you can work with other students to review your work.

CHM 101 PRACTICE TEST 1 Page 1 of 4

Solutions and Ions. Pure Substances

The Periodic Table of Elements

CHEM 171 EXAMINATION 1. October 9, Dr. Kimberly M. Broekemeier. NAME: Key

Speed of light c = m/s. x n e a x d x = 1. 2 n+1 a n π a. He Li Ne Na Ar K Ni 58.

Metallurgical Chemistry. An Audio Course for Students

(please print) (1) (18) H IIA IIIA IVA VA VIA VIIA He (2) (13) (14) (15) (16) (17)

Chem 102H Exam 2 - Spring 2005

PERIODIC TABLE OF THE ELEMENTS

single-layer transition metal dichalcogenides MC2

Secondary Support Pack. be introduced to some of the different elements within the periodic table;

Atomic Structure & Interatomic Bonding

M10/4/CHEMI/SPM/ENG/TZ2/XX+ CHEMISTRY. Wednesday 12 May 2010 (afternoon) 45 minutes INSTRUCTIONS TO CANDIDATES

Lab Day and Time: Instructions. 1. Do not open the exam until you are told to start.

INSTRUCTIONS: CHEM Exam I. September 13, 1994 Lab Section

If anything confuses you or is not clear, raise your hand and ask!

Chemistry 2 Exam Roane State Academic Festival. Name (print neatly) School

DO NOW: Retrieve your projects. We will be reviewing them again today. Textbook pg 23, answer questions 1-3. Use the section 1.2 to help you.

CMSC 313 Lecture 17 Postulates & Theorems of Boolean Algebra Semiconductors CMOS Logic Gates

CHEM 10123/10125, Exam 2

Chemistry 431 Practice Final Exam Fall Hours

Using the Periodic Table

7. Relax and do well.

Circle the letters only. NO ANSWERS in the Columns! (3 points each)

K. 27 Co. 28 Ni. 29 Cu Rb. 46 Pd. 45 Rh. 47 Ag Cs Ir. 78 Pt.

5 questions, 3 points each, 15 points total possible. 26 Fe Cu Ni Co Pd Ag Ru 101.

NAME: FIRST EXAMINATION

The Baltic Soil Survey

PHYSICAL SCIENCES MARCH CONTROLLED TEST GRADE

8. Relax and do well.

P a g e 5 1 of R e p o r t P B 4 / 0 9

UNIVERSITY OF CALGARY FACULTY OF SCIENCE MIDTERM EXAMINATION CHEMISTRY 353 READ ALL THE INSTRUCTIONS CAREFULLY

Lab Day and Time: Instructions. 1. Do not open the exam until you are told to start.

Chemistry Standard level Paper 1

CLASS TEST GRADE 11. PHYSICAL SCIENCES: CHEMISTRY Test 4: Matter and materials 1

Faculty of Natural and Agricultural Sciences Chemistry Department. Semester Test 1 MEMO. Analytical Chemistry CMY 283

Faculty of Natural and Agricultural Sciences Chemistry Department. Semester Test 1. Analytical Chemistry CMY 283. Time: 120 min Marks: 100 Pages: 6

610B Final Exam Cover Page

Connectivity in the LAC region

INSTRUCTIONS: Exam III. November 10, 1999 Lab Section

CHEM 107 (Spring-2004) Exam 2 (100 pts)

CHEMICAL COMPOUNDS MOLECULAR COMPOUNDS

Chem Exam 1. September 26, Dr. Susan E. Bates. Name 9:00 OR 10:00

CHEM 130 Exp. 8: Molecular Models

Essential Chemistry for Biology

8. Relax and do well.

IEC SYSTEM FOR CONFORMITY TESTING AND CERTIFICATION OF ELECTROTECHNICAL EQUIPMENT AND COMPONENTS

IEC SYSTEM OF CONFORMITY ASSESSMENT SCHEMES FOR ELECTROTECHNICAL EQUIPMENT AND COMPONENTS (IECEE)

Overview of numbers submitted for Statistics on Pending Mutual Agreement Procedures (MAPs) under the Arbitration Convention (AC) at the End of 2017

I M P O R T A N T S A F E T Y I N S T R U C T I O N S W h e n u s i n g t h i s e l e c t r o n i c d e v i c e, b a s i c p r e c a u t i o n s s h o

SCIENCE 1206 UNIT 2 CHEMISTRY. September 2017 November 2017

Country

CHEM 172 EXAMINATION 1. January 15, 2009

ORBITAL DIAGRAM - A graphical representation of the quantum number "map" of electrons around an atom.

PART 1 Introduction to Theory of Solids

Transcription:

Anycast atency How many sites are enough? Presented by Ricardo de Oliveira Schmidt October 25, 2016 Madrid, Spain Presentation copyright 2016 by Ricardo de Oliveira Schmidt

Reference: Anycast atency: How Many Sites Are Enough? Ricardo de O. Schmidt, John Heidemann and Jan Harm uipers Technical Report ISI-TR-2016-708, USC/Information Sciences Institute, May 2016 http://www.isi.edu/~johnh/papers/schmidt16a.pdf

Anycast Definition Multiple copies of a service at different locations configured to answer with the same IP address Purpose Distribution, redundancy, resilience, performance Where Content Delivery Networks (CDNs) DNS at the root and TD levels

Anycast

Anycast

Anycast

Anycast Anycast uses BGP to map users to sites But BGP only approximates closest match

RQ's and Approach If BGP only approximates closest in anycast: 1. How good (or bad) is this approximation? 2. What is the impact of location of anycast sites? 3. What is the impact of routing policy? Can we determine a minimum number of anycast sites as optimal? To answer these questions we...... study the relationship between latency and number of anycast sites... look at the Root DNS as case of study

Our case of study were four letters of the Root DNS: C: 8 sites : 58 sites : 33 sites : 144 sites We DO NOT focus on: What a letter should or not do Root DNS (or a part of it) C-Root -Root -Root -Root CC C C C C C C

RIPE Atlas Our vantage points (VPs) are RIPE Atlas probes 7,900 of them 174 countries 2927 ASes The bias towards Europe does not impact our qualitative results and conclusions

Measurements Step 1: determine catchment CHAOS queries to hostname.bind to anycast prefix...... returns a string that identifies a server CHAOS query anycast service S1 S2 AMS VP S1SYD.anycast-service.org S1 S2 SYD S1 S2 AX

Measurements Step 2: determine (optimal) latency ICMP (ping) to all sites of the anycast...... as given by the identifier from CHAOS anycast service RTT to AMS S1 S2 AMS VP RTT to SYD S1 S2 SYD RTT to AX Is RTT to SYD the lowest? S1 S2 AX

What Performance do we see? Distribution of RTT for all letters: 1 0.8 0.6 CD 0.4 0.2 0 median RTT ~30ms 0 50 100 150 200 250 300 350 RTT (ms) C-Root actual -Root actual -Root actual -Root actual We do not determine a threshold for good or bad performance...... but we do believe that 30ms is very good

What RTT is possible? 1 Comparing actual to optimal possible: C-Root with 8 sites is at its optimal -Root with 144 sites has plenty of room But their median is roughly the same CD 0.8 0.6 0.4 0.2 median RTT = 32ms 0 0 50 100 150 200 250 300 350 RTT (ms) C-Root C-Root actual C-Root optimal Note: Comparing anycast and unicast routing CD 1 0.8 0.6 0.4 -Root -Root actual -Root optimal 0.2 median RTT = 30ms 0 0 50 100 150 200 250 300 350 RTT (ms)

What about location? ocation matters! Simulation using C-Root sites 1 0.9 0.8 from west to east coast Note: CD 0.7 0.6 0.5 0.4 0.3 C-Root optimal AX+ORD+IAD+J AX+ORD+IAD AX+ORD AX (only) C-Root has sites in Europe Bias from RIPE Atlas is visible 0.2 0.1 0 0 50 100 150 200 250 300 350 RTT (ms)

What about location? ocation matters! Simulation using C-Root sites from US and Europe 1 0.9 0.8 0.7 0.6 CD 0.5 Note: C-Root has sites in Europe Bias from RIPE Atlas is visible 0.4 0.3 0.2 0.1 0 C-Root optimal CDG CDG + AX CDG + AX + J CDG + AX + J + RA 0 50 100 150 200 250 300 350 RTT (ms)

What about location? long distribution tail ocation matters! Simulation using C-Root sites from US and Europe 1 0.9 0.8 0.7 0.6 CD 0.5 Note: C-Root has sites in Europe Bias from RIPE Atlas is visible 0.4 0.3 0.2 0.1 0 C-Root optimal CDG CDG + AX CDG + AX + J CDG + AX + J + RA 0 50 100 150 200 250 300 350 RTT (ms)

Sites vs ocation So, at the end many sites do help those at the end of the distribution tail...... but it depends where and how connected RTT (ms) 300 200 100 A A A A A A A A A A A A A A A A A A A A A A A E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E S S S S N N N N N O O O O Africa () Asia (A) Europe (E) South America (S) North America (N) Oceania (O) C-Root 0 MY (20) SC (6) RE (5) MU (5) UG (5) TZ (5) ID (33) BD (13) NP (8) AU (69) CN (14) ZA (52) BT (5) IN (15) SG (45) PH (26) UY (10) C (16) E (10) NC (5) AR (18) P (7) H (22) SN (6) BR (49) NZ (53) GU (5) TW (5) (22) R (10) AE (8) JP (72) B (5) BJ (6) IR (28) AD (9) Z (21) CR (5) TN (7) I (18) CY (10) PR (6) GR (46) AM (17) GE (8) TR (20) IS (8) RU (307) BY (15) M (9) A (15) IE (75) BA (7) I (59) NO (83) T (10) V (15) EE (18) PT (48) RO (28) UA (128) SE (96) BG (66) IT (153) MD (8) P (92) RS (29) CA (132) HR (26) ES (93) BE (140) CZ (166) HU (36) DE (612) US (682) D (78) GB (347) CH (163) R (477) SI (28) AT (106) N (330) S (32) U (25) country code (# of VPs)

Sites vs ocation So, at the end many sites do help those at the end of the distribution tail...... but it depends where and how connected RTT (ms) 300 200 100 A A A A A A A A A A A A A A A A A A A A A A A E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E S S S S N N N N N O O O O Africa () Asia (A) Europe (E) South America (S) North America (N) Oceania (O) C-Root 0 MY (20) SC (6) RE (5) MU (5) UG (5) TZ (5) ID (33) BD (13) NP (8) AU (69) CN (14) ZA (52) BT (5) IN (15) SG (45) PH (26) UY (10) C (16) E (10) NC (5) AR (18) P (7) H (22) SN (6) BR (49) NZ (53) GU (5) TW (5) (22) R (10) AE (8) JP (72) B (5) BJ (6) IR (28) AD (9) Z (21) CR (5) TN (7) I (18) CY (10) PR (6) GR (46) AM (17) GE (8) TR (20) IS (8) RU (307) BY (15) M (9) A (15) IE (75) BA (7) I (59) NO (83) T (10) V (15) EE (18) PT (48) RO (28) UA (128) SE (96) BG (66) IT (153) MD (8) P (92) RS (29) CA (132) HR (26) ES (93) BE (140) CZ (166) HU (36) DE (612) US (682) D (78) GB (347) CH (163) R (477) SI (28) AT (106) N (330) S (32) U (25) country code (# of VPs) RTT (ms) 300 200 100 A A A A A A A A A A A A A A A A A A A A A A E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E S S S S S N N N N N N O O O Africa () Asia (A) Europe (E) South America (S) North America (N) Oceania (O) -Root 0 BD (14) NP (9) SC (6) IR (26) PH (20) PR (5) BT (6) SG (48) PT (52) SN (6) GE (11) JP (88) H (25) (20) BY (13) I (20) HR (37) BA (6) IQ (7) BJ (6) ES (112) A (17) IS (10) MY (20) MT (5) E (15) CY (13) DO (6) IT (189) M (13) EE (24) T (15) MD (9) RE (5) MU (5) B (6) PE (5) BG (70) N (407) HU (56) CA (150) IN (13) SI (34) P (120) S (36) DE (730) AT (147) RO (34) US (812) AD (8) MX (5) CZ (214) I (84) IE (88) U (37) GB (464) V (20) RU (371) ZA (64) P (7) AU (81) Z (21) BE (171) GR (62) UA (172) R (572) CH (191) TH (5) AR (20) MZ (5) D (98) ID (30) C (11) NZ (58) BR (42) GU (5) NO (107) AE (6) SE (135) R (14) UY (13) RS (34) TR (24) country code (# of VPs)

Sites vs ocation So, at the end many sites do help those at the end of the distribution tail...... but it depends where and how connected distribution tail RTT (ms) RTT (ms) 300 200 100 0 300 200 100 A A A A A A A A A A A A A A A A A A A A A A A E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E S S S S N N N N N O O O O MY (20) SC (6) RE (5) MU (5) UG (5) TZ (5) ID (33) BD (13) NP (8) AU (69) CN (14) ZA (52) BT (5) IN (15) SG (45) PH (26) UY (10) C (16) E (10) NC (5) AR (18) P (7) H (22) SN (6) BR (49) NZ (53) GU (5) TW (5) (22) R (10) AE (8) JP (72) B (5) BJ (6) IR (28) AD (9) Z (21) CR (5) TN (7) I (18) CY (10) PR (6) GR (46) AM (17) GE (8) TR (20) IS (8) RU (307) BY (15) M (9) A (15) IE (75) BA (7) I (59) NO (83) T (10) V (15) EE (18) PT (48) RO (28) UA (128) SE (96) BG (66) IT (153) MD (8) P (92) RS (29) CA (132) HR (26) ES (93) BE (140) CZ (166) HU (36) DE (612) US (682) D (78) GB (347) CH (163) R (477) SI (28) AT (106) N (330) S (32) U (25) country code (# of VPs) Africa () Asia (A) Africa () Asia (A) Europe (E) South America (S) Europe (E) South America (S) North America (N) Oceania (O) A A A A A A A A A A A A A A A A A A A A A A E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E S S S S S N N N N N N O O O North America (N) Oceania (O) C-Root -Root 0 BD (14) NP (9) SC (6) IR (26) PH (20) PR (5) BT (6) SG (48) PT (52) SN (6) GE (11) JP (88) H (25) (20) BY (13) I (20) HR (37) BA (6) IQ (7) BJ (6) ES (112) A (17) IS (10) MY (20) MT (5) E (15) CY (13) DO (6) IT (189) M (13) EE (24) T (15) MD (9) RE (5) MU (5) B (6) PE (5) BG (70) N (407) HU (56) CA (150) IN (13) SI (34) P (120) S (36) DE (730) AT (147) RO (34) US (812) AD (8) MX (5) CZ (214) I (84) IE (88) U (37) GB (464) V (20) RU (371) ZA (64) P (7) AU (81) Z (21) BE (171) GR (62) UA (172) R (572) CH (191) TH (5) AR (20) MZ (5) D (98) ID (30) C (11) NZ (58) BR (42) GU (5) NO (107) AE (6) SE (135) R (14) UY (13) RS (34) TR (24) country code (# of VPs)

inal Considerations ooking at real-world anycast deployments we learned that...... number of sites can make a difference on performance, but...... location and connectivity of sites have higher impact... we believe that ~12 well-connected sites is "enough" for performance or the future, we will focus on other purposes of anycast: Resilience to Denial of Service attacks oad balancing

inal Considerations ooking at real-world anycast deployments we learned that...... number of sites can make a difference on performance, but...... location and connectivity of sites have higher impact... we believe that ~12 well-connected sites is "enough" for performance or the future, we will focus on other purposes of anycast: Resilience to Denial of Service attacks oad balancing in about 20 minutes (depending on Wouter)

r.schmidt@utwente.nl http://www.ricardoschmidt.com/ Acknowledgements: Geoff Huston, George Michaelson, Ray Bellis, Cristian Hesselman, Benno Overeinder, Jaap Akkerhuis, Duane Wessels, Paul Vixie, Romeo Zwart, Anand Buddhdev, and C-Root operators. SIDN abs, Nnet abs and SURnet Self-managing Anycast Networks for the DNS (SAND) project http://www.sand-project.nl/ DNS Anycast Security (DAS) project http://www.das-project.nl/