forecast methodology – september 2015 - lightcounting methodology.pdf · forecast methodology –...
TRANSCRIPT
1 Market Forecast Methodology, September, 2015
Forecast Methodology – September 2015
LightCounting Market Research Notes
Vladimir Kozlov � September, 2015
2 Market Forecast Methodology, September, 2015
Summary In summary, the key assump=on of our forecast is that the growth rate of network bandwidth is correlated with growth in network traffic. Data on Internet traffic and the traffic inside datacenters (shown in slides 10-‐11) indicates that while the traffic con=nues to grow very quickly, the annual growth rate is slowing down. This trend is used as a key metric for calcula=ng the expected volume of DWDM ports and Ethernet transceivers in 2016-‐2020. Our forecast for the Ethernet transceiver market accounts for all Ethernet op=cal transceivers used in telecom, enterprise and datacenter applica=ons (slide 15). Slides 12-‐15 illustrate the disrup=ve effect of mega datacenters on the Ethernet transceiver market in 2010-‐2014. Our key forecast assump=on for the Ethernet market in 2016-‐2020 is shown as the blue line on slide 14. It projects that the annual growth rate of the combined bandwidth of Ethernet ports will decline gradually. It is very likely that we will see fluctua=ons in the actual annual growth rate in the next few years. Ini=al deployments of 100GbE may lead to another bump in the curve, similar to the peaks in 2010 and 2014, related to massive ini=al deployments of 10GbE and 40GbE, respec=vely. On the other hand, future economic downturns may lead to drops in the growth rate. Our forecast should average out and smooth over these fluctua=ons. This forecast model worked well in the past for established markets like DWDM (slides 6-‐7). However, it does not fully account for disrup=ons like op=cs usage in mega datacenters (slides 12-‐13). We look at the mega datacenter op=cs market as fairly well established by now, considering that it started back in 2007 with the ini=al deployments of 10GbE op=cs by Google, which were interrupted by the financial crisis of 2008-‐2009.
3 Market Forecast Methodology, September, 2015
Challenges to Forecasting • Limited historical data on
product sales • Uncertainty about the
future demand • Diverging expert opinions • A lot of hype in the media • Economic cycles • Market disruptions
4 Market Forecast Methodology, September, 2015
Key question: how fast will the market grow?
?
Forecast example
real market data
5 Market Forecast Methodology, September, 2015
LightCounting’s Approach • Data Mining and Validation
– Collecting data from more than 20 optical components and module vendors since 2004
– Cross-checking shipments of lasers, laser drivers and modulators with transponders, linecards or ports
• Model Development – Correlating sales of optical components and network
bandwidth to Internet traffic growth
• Model Validation – Critical reviews of previous forecast accuracy
6 Market Forecast Methodology, September, 2015
Growth rate of DWDM network bandwidth, calculated from port shipments
0%
10%
20%
30%
40%
50%
60%
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
Growth ra
te (%
)
CWDM / DWDM Internet Traffic
7 Market Forecast Methodology, September, 2015
Critical Review: Established Products 2.5G and 10G DWDM ports
-‐
100,000
200,000
300,000
400,000
500,000
600,000
2010 2011 2012 2013 2014
LightCoun3ng’s March 2010 Forecast
2.5 G 10 G
0
100,000
200,000
300,000
400,000
500,000
600,000
2010 2011 2012 2013 2014
Actual data
2.5 G 10 G
8 Market Forecast Methodology, September, 2015
Critical Review: New Products 40G and 100G DWDM ports
9 Market Forecast Methodology, September, 2015
August 2015 Forecast for 100/400G DWDM ports by application
-‐
50,000
100,000
150,000
200,000
250,000
300,000
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
Mega-‐DCI
Telecom
10 Market Forecast Methodology, September, 2015
Growth Rates in Aggregated Bandwidth
0%
20%
40%
60%
80%
100%
120%
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
Mega-‐DCI
Telecom
11 Market Forecast Methodology, September, 2015
Latest data on Internet traffic growth LINX Public Exchange Traffic
0%
10%
20%
30%
40%
50%
2010 2011 2012 2013 2014
Growth Rate
12 Market Forecast Methodology, September, 2015
Traffic growth rate at Google
Date Growth Rate
Jul-‐08 70%
Jun-‐09 47%
May-‐10 44%
Apr-‐11 39%
Mar-‐12 140%
Feb-‐13 75%
Dec-‐13 57%
Nov-‐14 52%
An increase of 50X over the period shown is a CAGR of 75%.
13 Market Forecast Methodology, September, 2015
1GbE and 10GbE Transceiver shipments
-‐
5,000,000
10,000,000
15,000,000
20,000,000
25,000,000
30,000,000
2010 2011 2012 2013 2014
LightCoun3ng’s March 2010 Forecast
1 GbE 10 GbE
0
5,000,000
10,000,000
15,000,000
20,000,000
25,000,000
30,000,000
2010 2011 2012 2013 2014
Actual data
1 GbE 10 GbE
14 Market Forecast Methodology, September, 2015
40GbE and 100GbE Transceiver Shipments
-‐
500,000
1,000,000
1,500,000
2,000,000
2010 2011 2012 2013 2014
LightCoun3ng’s March 2010 Forecast
40 GbE 100 GbE
0
500,000
1,000,000
1,500,000
2,000,000
2010 2011 2012 2013 2014
Actual data
40 GbE 100 GbE
15 Market Forecast Methodology, September, 2015
Growth rate in aggregated bandwidth of Ethernet ports
0%
10%
20%
30%
40%
50%
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
Growth ra
te (%
)
Ethernet Internet Traffic
16 Market Forecast Methodology, September, 2015
Our latest Ethernet forecast
$0
$1,000
$2,000
$3,000
$4,000
$5,000
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
Revenues -‐ total
DC Mega DC Enterprise Telecom