using this setup with a windows hyper v setup with pfsense. no issues

I am trying to determine if this model will meet our needs. We are looking at the HA Pair. Would be really powerfull as I see it right, the only thing is that actual at the moment the PPPoE Internet connection is only running on one CPU core and so far I don´t think that 1 GBit/s at the WAN port will be able to reach. But if you are connecting over a static public IP address all CPU core will be in usage and the 1 GBit/s is able to reach easy. Public facing website, with 2 providers each with 1G link. Normal traffic is around 100mbit/sec. 100 MBit/s would be not the problem!!! When we remove Cloudflare from the mix we push upwards of 600mbit/sec. We average around 150k states on the firewall. The 600 MBit/s would be also not the real problem as I see it right. The 150k states will be also reachable but this is more or less also pointed or related to the circumstance what on top is installed and activated on this units! Likes Squid & SquidGuard, Snort, pfBlocker-NG, HAVP & ClamAV or massively DPI tasks. Also have a site to site VPN that consumes about 25mbit/sec on average with peaks to 100mbit. If you use IPSec together with AES-GCM it will be using and benefiting from the AES-NI instruction set of the CPU SoC and later if Intel QuickAssist is also running or activated in pfSense the OpenVPN will be also profiting from this SoC. We are looking at moving to BGP for our two links. No problems here. OpenBGPD or Quagga will do the job likes you want. Currently running a single VM.  Sitting at 30mbit right now and here is the performance of the CPU: Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz Current: 324 MHz, Max: 2593 MHz 4 CPUs: 2 package(s) x 2 core(s) Load average  0.30, 0.27, 0.21 CPU 5% For sure a CPU core is not like the other and a high end Intel E5 Xeon CPU is not comparable to the Intel Atom C2x58 series. But for the named above things powerful enough as I see it right. I haven't been able to find much in regards to real performance on this hardware, any help would be appreciated. pfSense store SG-8860 1U HA system You can also easily call them directly or mail them a pre-sales question if you like it. US Phone: 1.512.900.2546 Email: sales@pfsense.org

@Sander88: Thanks for your reply. Yeah, I was thinking the same thing after reading some more on the internet… I have applied a fix (https://doc.pfsense.org/index.php/Tuning_and_Troubleshooting_Network_Cards#mbuf_.2F_nmbclusters) a few days ago. The NIC crashed 4 times in the past week, in one of these cases the mbuf graph show a full buffer (100% utilized). I have had no crashes since I applied this fix... so let's hope my PFSense box stays up and running ;-) Do you have the LCDproc dev package running ? Grtz DeLorean

Thank you for the reply, but i have bought a Cisco Console Cable from Ebay, and now it's working fine. pfSense 2.2.6 is installed en running on my XTM 510 ;-) Grtz DeLorean

i have the one with 4 onboard ports. here is how pfsense sees them skc0: <marvell gigabit="" ethernet="">port 0xd000-0xd0ff mem 0xec02c000-0xec02ffff irq 12 at device 0.0 on pci1 skc0: Marvell Yukon Lite Gigabit Ethernet rev. (0x9) sk0: <marvell semiconductor,="" inc.="" yukon="">on skc0 sk0: Ethernet address: xx:xx:xx:xx:xx:xx miibus0: <mii bus="">on sk0 e1000phy0: <marvell 88e1011="" gigabit="" phy="">PHY 0 on miibus0 e1000phy0:  none, 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-master, 1000baseT-FDX, 1000baseT-FDX-master, auto skc1: <marvell gigabit="" ethernet="">port 0xd400-0xd4ff mem 0xec020000-0xec023fff irq 10 at device 1.0 on pci1 skc1: Marvell Yukon Lite Gigabit Ethernet rev. (0x9) sk1: <marvell semiconductor,="" inc.="" yukon="">on skc1 sk1: Ethernet address: xx:xx:xx:xx:xx:xx miibus1: <mii bus="">on sk1 e1000phy1: <marvell 88e1011="" gigabit="" phy="">PHY 0 on miibus1 e1000phy1:  none, 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-master, 1000baseT-FDX, 1000baseT-FDX-master, auto skc2: <marvell gigabit="" ethernet="">port 0xd800-0xd8ff mem 0xec024000-0xec027fff irq 11 at device 2.0 on pci1 skc2: Marvell Yukon Lite Gigabit Ethernet rev. (0x9) sk2: <marvell semiconductor,="" inc.="" yukon="">on skc2 sk2: Ethernet address: xx:xx:xx:xx:xx:xx miibus2: <mii bus="">on sk2 e1000phy2: <marvell 88e1011="" gigabit="" phy="">PHY 0 on miibus2 e1000phy2:  none, 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-master, 1000baseT-FDX, 1000baseT-FDX-master, auto skc3: <marvell gigabit="" ethernet="">port 0xdc00-0xdcff mem 0xec028000-0xec02bfff irq 5 at device 3.0 on pci1 skc3: Marvell Yukon Lite Gigabit Ethernet rev. (0x9) sk3: <marvell semiconductor,="" inc.="" yukon="">on skc3 sk3: Ethernet address: xx:xx:xx:xx:xx:xx miibus3: <mii bus="">on sk3 e1000phy3: <marvell 88e1011="" gigabit="" phy="">PHY 0 on miibus3 e1000phy3:  none, 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-master, 1000baseT-FDX, 1000baseT-FDX-master, auto</marvell></mii></marvell></marvell></marvell></mii></marvell></marvell></marvell></mii></marvell></marvell></marvell></mii></marvell></marvell>

@oguruma: I'm not sure I understand. Do you mean that if I upgrade to the G3220 I will get 40% better power consumption and 2x performance or…..? That Pentium is faster than your current CPU and the whole system power consumption would go down ~40%

The 2.3 snapshots have been based on FreeBSD 10.3-PRERELEASE for a while now. the patch went in to STABLE-10 December 27 https://github.com/freebsd/freebsd/commit/b65a25360bf453761bef728656d2499296150eb6 So this has been fixed in 2.3 for quite a while now. All OP needs to do is load pfSense 2.3, and his NUC will work.

@BlueKobold: But then please make sure that the LEDs are really corresponding to the workload and not that they are not able to use by pfSense. The status LEDs do apply to connectivity on each interface, and there is a nice large 2x20 backlit LCD that tells me in percentage what the workload is, exactly how many kbit/mbit in and out, state table usage, CPU utilization, etc.  ;)

I would try adding '-d sat' to your smartctl command.  I use this to query the SMART data on my USB external hard drives. My command (under Linux) is something like: smartctl -a -d sat /dev/sdg There are other types you can pass with '-d', but this is the only one that's ever worked for me.

@coxhaus: I am new but interested. I am having a hard time deciding whether the OP is running the AT&T modem in bridge mode when he tests the pfSense build?  It is easy to tell.  If the pfSense build is getting an outside IP address.  Otherwise he has double NAT working against him. I just got TWC 300 megabit connection.  I am trying figure out how to maximize my connection speed. AT&T does not offer the ability to do a bridge mode, only the crappy IP Passthrough.

@duren: but how much less reliable than the Chinese 1037U's? Considering the NUCs are most likely made in Taiwan (I believe), what makes you think they are any less/more reliable? The one vbcn purchased has already got an issue.  ;D The 1037U on Aliepxress have Intel CPUs and Intel NICs supported by pfSense, the SDD and RAM you can buy locally if that helps your peace of mind (although they too are probably Chinese made…). And for all the posts in the Hardware/Vendors forum you are going to be hard pressed to find a single complaint about these boxes (from people who have actually bought and are using them). I'd recommend buying one and trying it, no VLANS needed. Or listen to BlueKobold and go buy a far more expensive Jetway box (which is also going to be made somewhere in asia...  ;D )

well darn. Wish I knew more about coding, I'd be tempted to investigate.

Are teh power savings of an mSATA or M.2 really worthwhile? Even if they are more energy efficient, does it really justify the cost given the higher cost to buy and replace them? No the cost is to high, but this was not your question, it was more the question if the SSD or mSATA is saving more electric energy compared against a spinning HDD. Also, where in Germany are you from/do you live? I was stationed in Heidelberg/Mannheim a few years ago, and I would kill to go back. My daughter was born in Schwetzingen. Lower Saxony, in the near region of Hannover.

I ran an optiplex 160 for about a year with version 2.2. It worked fine. Very easy to install and the performance was ok. You will need to configure vlan's and have a smart switch to create your WAN and LAN networks. Its not a bad little setup for home. The fanless design worked well and power usage was very low.

…just came back to add this: https://forum.pfsense.org/index.php?topic=87434.msg480143#msg480143 ...is also seen with the MS-9877 (also Atom D525 @ 1.80GHz). Temp of cores at 65°C (pfsense 2.2.6 64 bit full install with SSD wih TRIM, finally :-D ) With much older Atom Z510 had temps arround 45°C at exactly the same load (32 bit pfsense 2.2.6 full install with SSD). powerd enabled (tried minimum and hiadaptive yet). didn't buy much yet. waiting for doc to tell me the cooling is shyte... Or did I forget about some essential options to be set in the BIOS? Any suggestions highly appreciated!

yeah, I'm having the same thoughts.. Ok I'll get an unlocked Huawei modem today and try my luck the third time

So I fired it up. Sounds like it's about 35 to 40db. Not to bad if you're not in a quiet room but definitely makes a LOT of noise on startup. Sounds like any other server on startup. Screams like a 747, then idles down. Seems to idle at about 63Watts according to my watt meter. 63Watts isn't a lot but it isn't a little either and that's just with one power supply. Even with only one power supply installed it's 63 watts. Anyone know what the Cisco 300 idles at or any of the other suggested switches with 10Gb? BTW, I'm not trying to hijack the thread, just as curious as the creator of the thread as I'm also in the market for a decent switch but I don't want to spend 1,200 getting one either :P

Agreed. Unfortunately I need to get this one in place and have run out of time but I'll be building another one soon for CARP, that one will get further testing as this one will for now meet the needs of the two internet connections. One is 24Mbps the other 40 totally about 60Mbps. The reason I was trying to get more speed was for future reliability. So that if by chance the internet connections up here get faster and more stable this thing would be ready for the task and for general bench mark purposes. With this next one I'm going to do just as you said. Two machine, one on each side and then hammer traffic through it. First stock then with one package at a time. For those that were helpful, thank you. Some of you  ??? :-X