Forget about it. HFS+ and AFP are proprietary technologies made by Apple and not licensed to anyone.
Also keep in mind it’s a ROUTER - performance for file sharing will be horrible anyway.
HFS+ support in all Windows (and Linux) products came from reverse engineering. The best one from Paragon added journal replay which works only sometimes.
Speaking in terms of AFP all Linux products out there uses Netatalk implementation which is not as perfect as you may think. I’m running it for 3 years now, updated from sources every time it’s get updated - it works but it’s stability is discussable.
Speaking in terms of performance transferring at ~120Mb/s uses around 80% of one of two cpu cores on my home NAS which is x86 and have drives connected via SATA. Using USB drives will require more power.
Network storage is rather complicated area. If you don’t believe me look at commercial NAS solutions (keep in mind I’m not talking about pro-grade SANs) - every unit with decent speeds will run x86. I replaced my NAS with custom one because of two reasons: I had few ARM storages and they just sucks plus custom x86 was cheaper.
So yeah, I think router CPU is not going to handle network storage well - I did seen many with such functionality, none of them were near good. If you’re complaining about AEx speeds look how beefy ARM it’s holding - it’s 1Ghz + 512MB ram
for 10Gb ports im afraid even top CCR aren’t powerful enough, mostly because ROS didn’t scale much between many-cores, yet.
(unless you willing to sacrifice/“dumbify” config for speed, making its essentially “smart switch” for lan ports,which isn’t good idea for ANY router)
HFS+ and AFP not matter.
its like GOPHER in 2015 year. or http 0.9.
but 802.11ax would be cool, but such wireless chipsets isn’t come out yet and generally 60GHz support was quite rare at all
its meant “USE/configure as smart switch” ie with sacrificed/crippled/reduced config.
and no, CCR1009 do use switch chip. (and its ironically aged 8327 chip. even my hEX had 8337 as interface/PHY. which way, WAY better) http://i.mt.lv/routerboard/files/CCR1009-140630151432.pdf
“something” is always possible. question only if/who/when will do that.
And so, distracted by technicalities, everyone seems to have forgotten that RB3011 should have been available for a month already, which isn’t. So I’m wondering, is there something really wrong with the new to be product and should I go and buy something else?
I have learnt that it is normal when mikrotik doesn’t fulfill it’s timing given in announcements. Nothing special. Just take it as it is. The official answer was once to me: “It will be when it will be.”
Yes, it has the hardware, but no it’s not yet working - they haven’t gotten the drivers for it integrated into RouterOS yet. I presume that will be something that happens in ROSv7… whenever that might be. (It’s like sasquatch - there have been a few blurry pictures of it, but nothing concrete).
Does anyone perhaps have any numbers (or could run a quick benchmark) on what it can handle in terms of unaccelerated IPsec/AES? If the RB2011 could hit around ~40mbps, it shouldn’t be a stretch for the significantly faster processor in the RB3011 to get close to 100mbps at least?
i think hardware - advance faster than sofware implmentation for “advertised as hardware implementation” offloading of some portions of it.
A35, A32 cores, and even proposed mid tier A53 replacement (A72 Already get A73 to switch to and A32/A35 replace weaker ones)
and 1004k and his new, improved successor - make even MIPS-BE look not so bleak, perhaps.
in any/each case - cost and time of design/development was anyway remain significant to be problem. thats one of major reasons why ARM - outperform MIPS, PPC and x86 - they had notably quicker “time to market” performance, despite sometimes idential launch time for similar ideas in.