estimates.md

Power Estimates

Assumptions

• APs don't get to do a lot of power saving due to lots of mcast traffic
• High load conditions will mostly be during daytime (6am - 9pm) (assume full load on APs)
• At night APs are pretty much idle
• Offloader is constantly processing fastd traffic
• Rocket M2 is compareable to CPE210 @ 20 dBm
• PowerBeam is compareable to CPE510 @ 23 dBm
• 11 W Power Consumption from TL-SG3210 datasheet is accurate (switch)

Estimations

Daytime

Device	Count	Power
PowerBeam ISO 300	3	12.24 W
Rocket M2	4	12.80 W
Edge Point R6	2	5.28 W
Server	1	~40 W
Switch	1	11 W

Total: 81.32 W

Nighttime

Device	Count	Power
PowerBeam ISO 300	3	10.32 W
Rocket M2	4	12.16 W
Edge Point R6	2	5.28 W
Server	1	~40 W
Switch	1	11 W

Total: 78.76 W

measurements.md

Power measurements

CPEs

Load	CPE210 20dBm	CPE210 27 dBm	CPE510 23 dBm
Idle (ffki)	3.04 W	3.84 W	3.44 W
Full Load	3.20 W	4.72 W	4.08 W

Server

Setup

Tower server with

Key	Value
CPU	Intel i3-4150T
NIC	Generic Intel Gbit PCIe NIC
Storage	2x Samsung 120GB SSD (RAID1)
Fans	1x 80 mm, 2x 120 mm
PSU	300 W 80+ Bronze

Results

Load	Power
Idle	29 W
100%	41 W
200%	48 W
300%	49 W
400%	49 W
Routing 1 Gb/s	34 W
Fastd 35 Mb/s	39 W

EdgePoint EP-R6

Setup idle

EdgePoint with no devices connected

Load	Power
Idle	1.60 W

Setup ffki

EdgePoint with two CPEs, an offloader and a ThinkPad connected, switching only

Load	Power
Idle (ffki)	2.64 W
100 Mbit/s	2.64 W
1 Gbit/s	2.64 W

EdgePoint EP-R6 with CPE210 @ 20 dBm + CPE510 @ 23 dBm

Setup

4 ports on EdgePoint up, two of them with 24V passive POE

Load	Power
Idle (ffki)	7.54 W
50 Mbit/s (WLAN)	10.48 W

Expectation

To verify the above results in idle and full load condition the sum of the measured power consumption on the single devices must match the consumption of this setup

Idle

Device	Power
EdgePoint	2.64 W
CPE210	3.04 W
CPE510	3.44 W

Expectation: 9.12 W Measurement: 7.54 W

Full load

Device	Power
EdgePoint	2.64 W
CPE210	3.20 W
CPE510	4.08 W

Expectation: 9.92 W Measurement: 10.48 W

Conclusion

The power expectation matches the measured result pretty closely under high load conditions. In idle state the expected power consumption is quite a bit higher than the actual measured consumption. The discrepancy might be caused by unknowns like increased wireless activity during the first test of the CPEs alone. Anyway, the estimates seem to represent a safe upper limit for the power consumption of the whole setup.

Mercator Stromverbrauch Überschlag (max)

3x PowerBeam ISO 300:			 18W
4x Rocket M2:				 26W
2x Edge Point R6:			 14W
1x Intel i3 6100T			 35W
1x Server (Lüfter, Netzwerkkarten etc)	 20W
1x Managed Switch (TL-SG3210 o.ä.)	 11W
============================================
Summe:					124W