I’ve been checking https://wiki.mikrotik.com/wiki/MikroTik_News which has all the product releases available. I noticed that the RB4011 is all the way from 2018, the RB3011 2015, and the RB2011 2013. Does this mean that we’ll get the RB5011 this year? The biggest features that would be great are having switch hw-offloading with RSTP and maybe a more powerful CPU.
2 x SFP+ ports and poe on all ether ports
Keep dreaming
Mikrotik products always lack certain features.RB4011 swtich no vlan offload.CCR2004 no switch chip.RB3011 is too big.HAP AC2 lack memory.Which make them not perfect.You need to combine more products to achive your need.Mikrotik knows what you want, but just won’t give it to you
CCR2004 is a proper router and thus does not lack switch chip. The rest of devices on your list are SoHo devices (a completely different device group).
You’re saying RB3011 doesn’t fit standard 19" rack?
[/quote]
CCR2004 is a proper router and thus does not lack switch chip. The rest of devices on your list are SoHo devices (a completely different device group).
You’re saying RB3011 doesn’t fit standard 19" rack?
[/quote]
Who stipulated that there should be no switch chip in the router? The main problem is, If there is switch in CCR2004 people won’t buy CRS309 or such devices. And there is also no clear boundary to define Professional/SOHO/HOME router.
RB3011 lacks a shell of the same size as RB2011.It is too big for home use. There is a QCA switch chip with VLAN offload in RB3011 and was replaced by RTL in RB4011.
Mikrotik made a cage for himself. Also Trying to frame the user with this cage.
There is absolutely no need to combine router with the switch for devices of this grade.
For routing you buy a router, for switching - a switch (or multiple ones).
CCR2004 is capable of routing the network with tens of switches, and hundreds of devices.
All-in-one boxes are good for simple home networks, not for more complex ones.
If you add capabilities of CRS309 to CCR2004 it would cost nearly as much as two separate devices, and with far less flexibility that this solution offers, it will be a complete marketing disaster.
There is some “prosumers” that have a taste super all-in-one boxes (12 antennas with 802.11ax, 10G switch, PoE, a lot of CPU cores, etc..) Any professional knows this is the wrong way to do it.
That said, I also think that Mikrotik does some strange decisions sometimes.
“HAP AC2 lack memory”
What you need more memory for on that tiny thing?
DDR3 memory is really cheap today. Currently released WIFI wave 2 driver only works for ARM devices with more than 256 MB of RAM. It will take MT a lot of time to modify the driver to fit that tiny RAM. Time to market is the core advantage of a company’s products. Don’t waste of time. Others are working on wifi6 and wifi6e. Maybe WIFI7 is coming soon.
Mikrotik is still living in the past. There are no signs of them taking wifi6 seriously. By the speed of how things are advancing entire wifi2 package seems to be just side project of one of the developers, probably working on it just on weekends. Pretty much nothing have changed in the past year and there is no sign of any new hardware or any wifi improvements.
If you need your “RB5011” then either look around and see if some available devices may do what you need or just don’t buy Mikrotik at all. Waiting for new magic unicorn Mikrotik router to appear out of nowhere is just wishful thinking at this point. There are no plans, no roadmaps, no announcements of new devices, no new FCC listed devices…
What are the alternatives? The Ubiquiti Edgerouters lack features while being slightly slower and higher priced, TP-Link Omada is junk, Cisco co$t$ thou$and$, and making a Pf/OPNsense thing would break the bank compared to the RB4011’s cool $150 pricetag.
Guess what? Performance doesn’t come for free.
If you need performance because you have high speed WAN link for which you spend considerable amount of money, then spend some on decent router as well. If you have high speed WAN link which doesn’t cost much, then you surely have some budget to spend. If your WAN is not that fast, then a tad slower (and cheaper) router would do. If you need multi-gigabit routing betwern LAN segments, then reconsider LAN segmentation (or spend some money on essential infrastructure).
If you don’t want to spend decent money on network gear, then … well …
I just want the features that the RB4011 has (compact size, no-fuss OS system) but I don’t want it 3 years outdated.
Which part of RB4011 is 3 years outdated? The great thing about Mikrotik devices is that they come with insanely long support time. The only thing that outdates Mikrotik devices is lack of performance, other vendors tend to limit support to much shorter time and then device owner remains with devices running buggy software.
The processor could be a lot better. It uses Cortex A15 CPUs, which were released way back in 2012, 9 years ago!
For the record, RB4011 uses SoC AL21400 (SoC among other things features ARM cores but contains much more). This SoC can route around 2.5Gbps (give or take), IMO plenty for SOHO users now and good enough for vast majority in next few years.
If you trip on “features”, like CPU names, then you’re probably looking at wrong vendor. MT is essentially budget vendor which will likely use older SoCs because their price already dropped (because newer SoC was launched).
If you’re after “lot better processor”, then you can build a pimped-up PC and run CHR (or even ROS x86).
Xiaomi released a new WIFI 6 product AX9000 based on IPQ8072A which only costs 152.6472 dollar in China. So new chip solution does not mean expensive. Cortex-A15 uses 28nm process and Cortex-A72 uses 16nm. That means Cortex-A72 is more powerful and costs less. Cortex-A15 is a 32-bit core and it is really old. I can not find any new products based on it. The reason RB4011 uses A15 is for product division, otherwise it will lead to competition for high-end products such as CCR2004. MT’s products are not competing with other companies but competing with their own products. For ROS X86, MT said that fast path is a feature of Linux kernel. But why can’t it be implement on x86 ? I known there is no HWNAT in Tilera chipset but It indeed have fast path. So the reason is they can’t or just they don’t want to? MT also said that the drivers for X86-64 are different and cannot be used in ROS 6. Enabling x86-64 is very difficult. But x86-64 can be found on CHR. And i have run x86-64 on bare machine for many years.
Fast path means nothing to me anyway, since to use queues you need to disable it.
You known, as lack of some key features such as Fast-path & Fast-track. X86 is treated as a second-line product by MT. It can’t even support the old platform like J1900 well (USB not working). So the “insanely long support time” is a joke.