📜 ⬆️ ⬇️

Study of the sustainability of national segments of the network


Top 20 sustainable regions for 2017 on the world map

The Internet as a system of networks that interact through the nodes of telecom operators (autonomous systems - AU), the more stable the more alternative routes between the AUs, the simple principle of fault tolerance. This study demonstrates how the failure of a single but most significant operator in the region affects the global availability of the national segment.

The global availability of any telecom operator is affected by its paths to Tier-1 operators. Tier-1 - transnational / transcontinental operators providing global connectivity between continents and countries.
')
If there are no paths to Tier-1 operators, the operator will not have global accessibility.

Suppose that a given operator is experiencing problems, and, accordingly, the passage of traffic through it is difficult or impossible. We decided to find the answer to the question: “What percentage of the remaining telecom operators in the region will lose indirect communication with Tier-1 operators, and hence global accessibility?”.

Why do we simulate such a situation? Strictly speaking, when modern Internet was conceived, it was assumed that each operator would have at least two independent upstream vendors, which guarantees fault tolerance in case of problems of one of them. However, for many networks, this remained a recommendation on paper. Therefore, the failure of a single operator can lead to systemic problems for a whole set of downstream telecom operators, even to the inaccessibility of entire national segments. But do major transit operators fail in the real world? The answer is yes, and with enviable regularity. And if there has not been a single gap, then it is worth remembering the law of Murphy: "What can happen will happen."

The modeling process was divided into the following steps:

  1. We built alternative routes to Tier-1 operators for each speaker in the world using the model of relations between speakers of the Qrator.Radar project;
  2. Using the Maxmind geobase, we compared the address space of the AU to different countries;
  3. For each speaker, we normalized the received geodata to avoid a situation where the speaker has a degenerate presence in the region. An example of such a degenerate presence is Hong Kong, where there are hundreds of operators on the largest Asian Internet Exchange, HKIX, and they have nothing to do with accessibility in Hong Kong itself;
  4. Further, the impact of the hypothetical failure of a given AS on other communication operators, and as a result, on the national segments was assessed;
  5. For each country, the AU was allocated, the effect of the failure of which will have the greatest impact on the other AU of the country.

Below is a table of the top 20 sustainable countries in 2017 and recalculated according to the new top 20 for 2016.
% failure
2016 year
Change 2016 - 2017
2017 year
% failure
2.57478
Germany (DE)
1 place
Germany (DE)
2.29696
3.14068
Canada (CA)
Down 2 positions
Hong Kong (HK)
2.65659
3.46469
Switzerland (CH)
3rd place
Switzerland (CH)
3.57245
4.03446
United Kingdom (GB)
Down 2 positions
Canada (CA)
3.67367
4.19754
Hong Kong (HK)
Up 3 positions
France (FR)
3.68254
4.34753
Ukraine (UA)
Down 2 positions
United Kingdom (GB)
3.76297
4.39691
United States (US)
Down 2 positions
Belgium (BE)
3.93768
4.83975
Belgium (BE)
Up 1 position
Ukraine (UA)
3.95098
5.68121
Spain (ES)
Down 8 positions
United States (US)
3.97103
5.78643
Poland (PL)
Down 6 positions
Bangladesh (BD)
5.29293
5.99955
France (FR)
Up on 6 positions
Romania (RO)
5.35451
6.00547
Russia (RU)
Down 1 position
Brazil (BR) - newcomer
5.39138
6.39252
Australia (AU)
Departure from the top 20
Russia (RU)
5.73432
6.88687
Ireland (IE)
14th place
Ireland (IE)
5.87254
7.0508
Romania (RO)
Up on 4 positions
Czech Republic (CZ) - Newbie
5.88389
7.43945
Austria (AT)
Down by 3 positions
Poland (PL)
5.99655
7.84456
Italy (IT)
Departure from the top 20
Bulgaria (BG)
6.20975
7.97141
Bangladesh (BD)
Up 8 positions
Spain (ES)
6.58064
8.14681
Bulgaria (BG)
Up 2 positions
Austria (AT)
7.14221
8.15989
Philippines (PH)
Departure from the top 20
Luxembourg (LU) - newcomer
7.28208
As you can see, there are several permutations in this list, but in this case it cannot be said that with a difference of tenths of a percent, there is some fundamental change in quality. Speaking about the top 20, and about other countries, the refusal in which a single operator leads to the inaccessibility of no more than 10% of the regional AU (29 such countries) - all these countries have a fairly diversified market for IP transit services, with a large number of alternative routes .

Full table
Country code
AC number
% failure
DE
8881
2.29696
HK
3356
2.65659
CH
3303
3.57245
CA
577
3.67367
FR
174
3.68254
GB
174
3.76297
BE
6848
3.93768
UA
3255
3.95098
US
174
3.97103
Bd
58587
5.29293
RO
9050
5.35451
BR
12956
5.39138
Ru
12389
5.73432
IE
174
5.87254
CZ
39392
5.88389
PL
12741
5.99655
Bg
8717
6.20975
ES
3352
6.58064
AT
1764
7.14221
LU
6661
7.28208
IT
12874
7.86486
JP
2914
7.97843
SG
4657
8.85406
BA
9146
8.91981
TH
7470
9.45542
MX
8151
9.65379
AU
1221
9.78872
SE
12552
9.84424
SI
5603
10.1375
VN
24088
10.2254
PH
4775
10.2797
PR
23520
10.4444
KG
8449
11.1033
PE
28032
11.1842
NO
2116
11.521
Ug
37100
11.6839
CD
12491
11.7453
Bz
263824
12.2882
AL
21183
12.4086
HU
12301
12.7217
GR
5408
12.7795
ZA
37100
12.8649
Nz
45177
12.9262
Ke
33771
13.0855
CO
23520
13.1462
My
38182
13.3225
KZ
41798
13.4615
CR
23520
13.6906
Ni
262206
13.9167
TT
5639
13.933
AR
12956
13.9918
MT
15735
14.0475
IM
13122
14.1363
LI
12387
14.2122
SD
36998
14.2594
HR
35549
14.2674
SS
37376
14.6274
KR
9318
15.1173
RS
8400
15.3291
PK
38193
15.3924
CY
8544
15.6266
LK
9329
15.8408
Mk
43612
15.8933
SA
39386
16.1392
ID
7713
16.3255
NE
37385
16.5226
MD
35346
16.6024
Dz
36947
16.6246
SK
5578
16.658
EE
3327
16.6605
IL
8551
16.7245
Mu
3741
16.9752
IN
9498
17.2793
LT
8764
17.4499
CL
14259
05.0572
EG
8452
12/18/29
HN
262206
18.4023
Bh
41426
18.5365
Rw
37228
18.6543
Tw
9505
18.6924
Bf
37282
18.9097
Tj
8449
18.9292
SC
36958
18.9609
SM
15433
19.0898
Gh
29614
19.1502
Ge
35805
19.1554
AO
37468
19.4097
Mw
37440
19.4405
IS
44735
20.0402
LV
12578
20.2936
Dk
31027
20.8775
EC
19169
21.3586
SV
262206
21.4237
MA
6713
21.516
GP
17149
21.5597
TR
15924
22.1705
PT
8657
22.2313
FI
47605
22.2788
Iq
44217
22.5989
BO
12956
23.0191
BM
3855
23.2038
Mz
37100
23.5821
CN
4134
23.7955
Tz
37084
23.8041
Zw
37204
23.8109
AM
49800
24.0208
NG
37282
24.056
GT
12956
24.3309
BS
15146
24.5646
ZM
30844
24.5807
VA
61160
24.6227
KH
24492
24.985
Fj
45349
25.4647
JO
8697
26.9172
VG
22069
27.4085
GI
34803
27.5376
BY
12406
27.684
VI
20080
27.9791
NP
45845
28.3101
Bj
28683
28.5238
KW
43852
28.708
GM
5511
28.7522
Lr
16637
28.7881
KY
3855
28.7942
OM
8529
29.0397
GN
60171
29.5369
PA
23520
29.5604
Py
23201
29.8865
GU
9246
29.9052
CM
15964
29.9918
CG
37451
31.2056
UY
6057
31.2855
MM
131322
32.4047
XK
33983
32.6934
BI
37545
33.0276
Fo
206928
33.2005
Tg
24691
33.2005
SN
37196
33.4426
CI
29571
35.144
Do
23520
35.8244
GA
31167
36.3772
AG
23520
36.8915
Gg
34497
36.9249
NA
33763
37.3629
ME
15958
38.1412
PS
12975
38.7881
LS
37057
38.8161
Bq
52389
38.8267
Td
60725
39.3523
MG
37054
39.9269
Je
13208
40.2594
BB
23520
40.8275
AE
8966
41.3346
BL
21538
41.8989
Ye
12486
42.6058
Gf
5511
42.6892
Mq
48252
43.5538
Mf
33392
43.6176
LA
24337
44.3725
RE
37002
44.4138
CW
11081
44.5206
CF
37460
45.6704
QA
8781
48.7885
VE
8048
49.2335
Ly
21003
49.6574
SR
52398
49.7463
MR
37541
49.7884
HT
27759
49.7959
Gy
19863
49.843
BT
38740
49.9102
DM
23520
50.75
AZ
29049
54.9575
Bw
37678
57.0625
SL
60171
57.1429
WS
38227
58.846
Lb
42020
59.1988
IR
12880
59.4523
AI
52389
60.3524
TN
2609
62.3294
ML
30985
63.1634
MV
7642
63.2251
Tk
57382
64.6496
AF
55330
65.52
Tm
20661
66.2846
Ax
200134
66.3124
MN
45204
66.496
VU
45495
66.538
PF
55943
66.6533
SO
37662
67.4491
SB
38456
67.7938
MS
54334
69.8795
Sx
11081
70.8825
ST
6774
76.7106
Gq
60171
77.3256
Sz
19711
77.6899
AD
6752
82.291
TO
54334
83.3333
Ck
10131
83.8574
KN
36290
85.8369
PG
17828
87.222
MO
4609
87.3973
NU
54334
87.5
SY
47447
88.75
DJ
30990
89.251
Jm
23520
91.7012
KP
4837
92.2169
AW
11816
93.0463
KM
36939
94.1974
Tl
7713
95.1494
CV
37517
95.4168
Yt
37662
95.4914
GW
327769
96.5158
TC
23520
96.9121
Bn
55303
97.0685
VC
23520
97.6697
LC
23520
97.7676
NR
55723
97.7899
AS
4323
98.4665
GD
23520
98.7318
GL
8818
98.8338
Pw
4648
98.8338
ET
24757
98.8338
PM
11260
98.8435
KI
134783
98.8533
AQ
204136
99.5405
CU
11960
99.6033
ER
30987
99.6612
MC
6758
99.7771
MP
7131
99.8242
NC
18200
99.8371
Uz
28910
99.9425
MH
6453
99.98
FM
10130
99.98
GS
54334
100
Io
44972
100
PN
54334
100
SH
54334
100
Sj
54334
100
FK
54334
100
Nf
54334
100
CC
394021
100
CX
54334
100
Tv
11029
100


Separately, it is worth noting the strong influence of AS 174, the Cogent operator, on a whole range of European national segments: France, the United Kingdom, the United States, and Ireland only in the top 20 sustainable regions. That is, the occurrence of problems in the autonomous system 174 may lead to problems in the group of neighboring regions. Although not complete inaccessibility, as we are talking about well-developed and actively decentralized national networks.

In Russia, there are no surprises. Rostelecom remains the backbone operator, but even its hypothetical offline will result in inaccessibility of no more than 5.73% of Russian telecom operators. At the same time, the percentage of refusals, comparable to last year’s, provides Russia with only 13th place in the ranking. There are practically no networks of international Tier-1 operators in Russia, however, the domestic market is represented by a whole set of large and medium Tier-2 networks, which ensures such high stability.

Is the largest operator in the region always the greatest influence on the availability of other operators? As shown by our calculations - not necessarily. For example, in Germany, the largest telecom operator is Deutsche Telecom, but from the point of view of global connectivity, the failure of AS8881, owned by Versatel, will have the greatest impact on the AU in the region. And according to our estimates, the trend towards the increasing importance of Tier-2 operators will continue in the near future.

Speaking about global trends, the average “instability” in 2017 was 41%, which is 1.6% lower than a year earlier.


Gradient from dark green (0%), through yellow (25%) to red (50%) color, indicating the degree of dependence of the segment from a single supplier.

The greatest positive changes for the year occurred in the developing market of Africa. Thus, sustainability in several countries in the region, including the Gambia and Liberia, has increased by almost 40%. However, the movement towards increasing the stability of national segments cannot be called unidirectional. For example, the dependence of the national segment of Jamaica to the failure of a single operator has increased from 34% to 91%, the external connectivity of this country is now almost entirely dependent on the work of the AC 23520 (Columbus Networks).


Regions with 99% or more dependency on a single supplier.

At the very bottom of our rating there are 10 territories, which in case of failure of a single operator / autonomous system will experience 100% inaccessibility from the point of view of the rest of the world. Of the countries of the post-Soviet space, Uzbekistan is closest to them, located on 235 place with a value of 99.94% of inaccessibility due to the failure of AS 28910. The rest of the countries are island microstates.

In addition to the countries already mentioned, Cuba and South Sudan are the most well-known reader countries with more than 99% dependence on a single communications provider. North Korea has 92%. It is curious that Monaco has very poor connectivity (the only red dot in Europe), while another micro-state, Luxembourg, closes the top 20 sustainable countries in 2017.

Against the background of growing cyber threats and more and more emerging news about attacks on the Internet infrastructure, governments, companies and users from countries where resiliency is less than 50%, and 25% in an amicable way, you should seriously think about your own situation. We have already seen large-scale attacks on the largest national providers and major service providers of connectivity and availability, such as DNS-providers, as well as the consequences of such attacks in terms of the availability of a network resource for a specific user from anywhere in the world.

The results of the study demonstrate that the market of telecom operators, built on the principle of competition, ultimately turns out to be much more resilient both to problems within and outside the region. In the opposite case, the disconnection of a single provider can lead to the unavailability of network resources from a significant proportion of users of the whole country or region.

Source: https://habr.com/ru/post/332844/


All Articles