broadband quality war: are you a winner or a...

35
Broadband Quality War: Are you a winner or a loser? IESL Auditorium, Colombo, Sri Lanka April 29, 2009 1

Upload: vuongquynh

Post on 22-May-2018

216 views

Category:

Documents


1 download

TRANSCRIPT

Broadband Quality War:

Are you a winner or a loser?

IESL Auditorium, Colombo, Sri Lanka

April 29, 2009

1

1. Trends identified from QoSE testing

2. Policy interventions based on research

2

Trends identified from QoSE testing

Fixed Broadband in Colombo

3

0

200

400

600

800

1,000

1,200

1,400

1,600

1,800

2,000

0800 1100 1500 1800 2000 2300

SLT (2 Mbps) Colombo - 2008 Dialog (2 Mbps) Colombo - 2008

Download speed: Feb 2008

Relatively healthy in ISP domain…

Time

Do

wn

loa

d s

pe

ed

(k

bp

s)

4

0

200

400

600

800

1,000

1,200

1,400

1,600

1,800

2,000

0800 1100 1500 1800 2000 2300

SLT (2 Mbps) Colombo - 2008 Dialog (2 Mbps) Colombo - 2008

Download speed: Feb 2008

… poor in local domain …

5

Time

Do

wn

loa

d s

pe

ed

(k

bp

s)

-

200

400

600

800

1,000

1,200

1,400

1,600

1,800

2,000

0800 1100 1500 1800 2000 2300

SLT (2 Mbps) Colombo - 2008 Dialog (2 Mbps) Colombo - 2008

Download speed: Feb 2008

… poorer when accessing INT sites

6

Time

Do

wn

loa

d s

pe

ed

(k

bp

s)

Download speed: From Feb 2008- Feb 2009

Not too different in ISP…

7

Time

Do

wn

loa

d s

pe

ed

(k

bp

s)

Download speed: From Feb 2008- Feb 2009

Improved in local…

0

200

400

600

800

1,000

1,200

1,400

1,600

1,800

2,000

0800 1100 1500 1800 2000 2300

SLT (2 Mbps) Colombo - 2008 Dialog (2 Mbps) Colombo - 2008

SLT (2 Mbps) Colombo - 2009 Dialog (2 Mbps) Colombo - 2009

Why?Why?

8

Time

Do

wn

loa

d s

pe

ed

(k

bp

s)

Download speed : From Feb 2008- Feb 2009

Improved in INT Why?Why?

9

Time

Do

wn

loa

d s

pe

ed

(k

bp

s)

Trends identified from QoSE testing

Fixed Broadband in Dhaka

10

Download speed: From Oct 2008- Feb 2009

Dropped in ISP

0

50

100

150

200

250

300

0800 1100 1500 1800 2000 2300

Sirius (256 kbps) Dhaka - 2008 SKYbd (256kbps) Dhaka - 2008

Sirius (256 kbps) Dhaka - 2009 SKYbd (256kbps) Dhaka - 2009

11

Time

Do

wn

loa

d s

pe

ed

(k

bp

s)

Download speed: From Oct 2008- Feb 2009

Dropped in INT

0

50

100

150

200

250

300

0800 1100 1500 1800 2000 2300

Sirius (256 kbps) Dhaka - 2008 SKYbd (256kbps) Dhaka - 2008

Sirius (256 kbps) Dhaka - 2009 SKYbd (256kbps) Dhaka - 2009

12

Time

Do

wn

loa

d s

pe

ed

(k

bp

s)

Trends identified from QoSE testing

Fixed Broadband in Chennai and New Delhi

13

Download speed: Feb 2008 – Feb 2009

Over-delivery in ISP

0%

20%

40%

60%

80%

100%

120%

140%

160%

180%

200%

0800 1100 1500 1800 2000 2300

BSNL (2 Mbps) Chennai - 2008 Airtel (512 kbps) Chennai - 2008

0%

20%

40%

60%

80%

100%

120%

140%

160%

180%

200%

0800 1100 1500 1800 2000 2300

Airtel (256 kbps) Delhi - 2009 MTNL (256 kbps) Delhi - 2009

Airtel (256 kbps) Chennai - 2009 BSNL (256 kbps) Chennai - 2009

Time

Act

ua

l a

s a

% o

f st

ate

d

Time

Act

ua

l a

s a

% o

f st

ate

d

14

% of Advertised in 2008 % of Advertised in 2009

Download speed: Feb 2008 – Feb 2009

Over delivery in INT

0%

20%

40%

60%

80%

100%

120%

140%

160%

180%

200%

8:00 11:00 15:00 18:00 20:00 23:00

BSNL (2 Mbps) Chennai - 2008 Airtel (512 kbps) Chennai - 2008

0%

20%

40%

60%

80%

100%

120%

140%

160%

180%

200%

0800 1100 1500 1800 2000 2300

Airtel (256 kbps) Delhi - 2009 MTNL (256 kbps) Delhi - 2009

Airtel (256 kbps) Chennai - 2009 BSNL (256 kbps) Chennai - 2009

15

Act

ua

l a

s a

% o

f st

ate

d

Act

ua

l a

s a

% o

f st

ate

d

% of Advertised in 2008 % of Advertised in 2009

Time Time

Trends identified from QoSE testing

‘The BIG picture’

Fixed Broadband in Chennai,

Colombo, Dhaka and New Delhi

16

Prices*

Mo

nth

ly r

en

tal

(US

D)

17* Monthly rental (without tax) of selected packages ; not a direct comparison as other

conditions apply.

* Monthly rental (without tax) of selected packages ; not a direct comparison as other

conditions apply.

Packages

Download speed: Feb 2009 - ISP

Sri Lanka: Good speeds but false advertising

Time

Do

wn

loa

d s

pe

ed

(k

bp

s)

Time

Act

ua

l a

s a

% o

f st

ate

d

Actual % of Advertised

18

Download speed: Feb 2009 - INT

Sri Lanka: Good speeds but false advertising

19

Time

Do

wn

loa

d s

pe

ed

(k

bp

s)

Time

Act

ua

l a

s a

% o

f st

ate

d

Actual % of Advertised

Return Trip Time to INT: Feb 2009

Yet to meet Singapore (IDA) standard

Time

ms

20

300 ms300 ms

Jitter – pinged to INT: Feb 2009

Almost within acceptable levels

21

Time

ms

50 ms 50 ms

Time

Packet Loss – pinged to INT: Feb 2009

Almost within acceptable limits

22

3%3%

Trends identified from QoSE testing

HSPA in Colombo

23

Prices*

24

* Monthly rental (without tax) of selected packages ; not a direct comparison as other

conditions apply.

* Monthly rental (without tax) of selected packages ; not a direct comparison as other

conditions apply.

Mo

nth

ly r

en

tal

(US

D)

Packages

Download speed: Feb 2009

ISP

25

Time

Do

wn

loa

d s

pe

ed

(k

bp

s)

Time

Act

ua

l a

s a

% o

f st

ate

d

Actual % of Advertised

Download speed: Feb 2009

INT

26

Time

Do

wn

loa

d s

pe

ed

(k

bp

s)

Time

Act

ua

l a

s a

% o

f st

ate

d

Actual % of Advertised

Return Trip Time to INT: Feb 2009

Yet to meet Singapore (IDA) standard

Time

ms

27

300 ms300 ms

Jitter – pinged to INT: Feb 2009

Not too different from fixed

Time

ms

28

50 ms50 ms

Packet Loss – pinged to INT: Feb 2009

Not too different from fixed

Time

%

29

3%3%

Policy interventions based on the research

30

31

Sri Lanka:

Response to

the paper on

‘NATIONAL

BACKBONE

NETWORK’

Sri Lanka: Response to paper on

‘NATIONAL BACKBONE NETWORK’

• A ‘National’ backbone will not solve all QoSE issues

• 2 Mbps is too low for a national backbone capacity

• Interconnection issues are crucial and must be

addressed

• Instead of uniform capacity everywhere the NBN,

should be designed to cater to specific demands to

provinces

32

Bangladesh: Response to the paper on

‘BROADBAND WIRELESS ACCESS SERVICES’

• Operators should guarantee QoSE not within ISP

only, but till first entry point to US

• Operators should publish contention ratios

• Assurance at launch is inadequate; QoSE should be

regularly monitored

• “Broadband = 128 kbps +” definition should change

33

India: Response to the paper on

‘ISP BANDWIDTH REQUIREMENTS’

• Suggested contention ratios 1:20 (business) and 1:

50 (residential)

• Information on contention ratios should be made

public

• Broadband QoSE is not just speed; need a holistic

view

• Operator obligation should be till first entry point to

US

– Adopted 1:30 and 1:50

34