Just to preface this we've deployed hundreds of the RB100 and RB500 series boards, and have been using mikrotik for years.

This abrupt change in footprint has left us wondering if we should continue to use mikrotik or switch to a platform that won't change at the drop of a hat.

The new boards are good, but the fact is now we have to go out and replace existing infrastructure and redesign our enclosures to support these new boards is really a pain in the ass. What would have taken a relatively short period of time to change out a board if it failed, now takes hours to replace the enclosure, and redo all the LMR. Some instances where we've used ARC panels with 133 boards for a small "repeater" is now impossible because a RB433 will not fit in the ARC enclosure. So we have to scrap that design all together and completely replace it...

One suggestion... Next time put yourselves in the position of your customers who have hundreds of these devices deployed. Do we really want to go out and have to redesign an entire solution to conform to your new standards? Or maybe.. ask those on the forums, you have an outlet right here!! Why not talk to some of your customers and say.. hey this is the direction we're thinking of going, do you see any problems with this?

I'm a little pissed off right now because I've had numerous 500/100 boards fail recently we've gone through our inventory and no supplier has any replacements...

So now I need to find a board similar to the 500's that will run RouterOS, start developing plans to redesign sites using the 500/100 boards, or bite the bullet and start deploying a different solution..

It would also be nice to see support for Intel's XScale architecture because there are products available already that have similar dimensions to the 500 series and support up to 4 miniPCI cards.. But I'm sure that would impact sales of the Routerboards you currently produce...

I couldn't agree with you more, after deploying hundreds of RB112's and RB532's I too am really frustrated with the change to the new layout. I am hoping that they learned their lesson and continue to roll out any new products with the same sizing as the RB133, RB411, and RB433 series so I don't have to go through this fiasco again.

However, I must say that I didn't mind the change in mounting holes as much as the idiotic triple-shot m-pci layout. All they would have to do is take the bottom m-pci connector in the triple-shot layout put it on the bottom facing the other direction, and then lower the middle triple-sot m-pci connector on the top and you'd have a much more usable board. You could easily fit 3 x XR5 cards, and the board size would fit in 95% of all the enclosures on the market with cards installed.

I really like RouterOS, but between the v3.x not working well on the RB1xx series and the board layout change I have been thinking if Mikrotik is still the best choice for me in the long haul.

Was just thinking and it would be nice to here a Mikrotik representive speak on this issue. I would have much more confidence in Mikrotik if they pledged to stick with a layout for 'x' years and will provide 'x' number of years overlap if changes are planned. Take it from the computer industry, to stray from a well entrenched standard is not easy and/or good for business just compare BTX to ATX, Intel long long ago thought BTX would overtake ATX, and yet it still isn't even close.

Cheers

ummm, where were you a few months ago? this topic is already settled for a while, because all new boards come with the same mounting hole locations, and the same width.

as far as i know RB133/RB133c have same mount-holes as RB333 (already outdated) RB411 RB433 RB433AH

so as far as I can tell - there is no problem. if you are complaning about Rb532, then you are a few years too late

Don't agree Normis. It IS an ongoing problem and its not too late to fix it.

We have at least 250 RB532 plus several hundred more 100 series boards in the air that gradually fail. Its going to take years to need to replace all of them. And every time we do need to replace one, we need to replace the PacWireless outdoor enclosure they are sitting in. Do you have any idea how much money this is costing your customers?? We are at the point now where hardware costs are so low they are largely irrelevant, the real costs of doing business are labour and lost opportunity cost, when we could and should be out doing something constructive rather than backtracking because MikroTik changed a form factor.

It seems from the comments above that MikroTik never took the replacement market into consideration, which is pretty surprising.

If you were to come out with a new fast Routerboard that used the 532 form factor we would be pathetically grateful. We think of MikroTik every time we have to replace an enclosure and all its cabling at up to 100M in the air in the frozen north in winter...

George

Listen to what you are saying!

What about all the users of the mentioned devices, who would then have the same comment as the original poster?

We have 14 different devices sharing the same mounting hole locations and width, which means that we have "found a form factor and stuck with it" and you are suggesting we now ruin the form factor just to help you migrate a few year old product, which is the only one with those dimensions?

I understand the issue with RB532, but we have fixed it with new products, and their sizes are now the same.

I am not saying you trash your existing product line, exactly the opposite. Keep the new form factor for as long as you can, and provide notice of end of life well before you decide to do away with the mainstream products in the current form factor.

I am asking you to make ONE Routerboard that has a modern processor and fits the 532 form factor. There is a decent market for this product that will be there for years and years to come. Perhaps a poll might be in order to see how big such a market might be and whether this would be worth considering?

MT is hurting their oldest and largest customers. That generates a trust issue as has been mentioned elsewhere on this thread. Forgive me for blowing off some steam over this, but it really is becoming a big problem and we are trying to figure out how to deal with it. Eighty-five towers with an average of eight radios per tower. Easy to see the size of the issue. I have one hundred-metre tower with a dozen enclosures on it, all of which were designed to accomodate a 532 and all of which are going to have to be changed out.

We are now at the point where we are seriously considering pulling radios down off the towers and replacing Ethernet with either hardline or waveguide risers in order to cut back on maintenance overheads. We use a lot of waveguide already so we have the experience in house.

The costs of replacing all existing enclosures to fit the new form factor, plus the extended downtime associated with repairs just about justifies the cost differential in running hardline, and once we've taken the hit in moving to the tower base ongoing maintenance becomes a snap.

Backward compatibility is a big deal. In any and all industries, but particularly communications where truck rolls to remote tower sites are very expensive. You already know why they make Long Term Support versions of Linux of course... We need a Long Term Support version of an RB Access Point, at least regarding the form factor, number and location of slots and number and location of Ethernet ports. We thought you understood that judging by all the improved versions of the 532 that were released over time.

By all means change the processor and anything else necessary as technology changes. If you consider that anyone that bought large quantities of earlier products is just out of luck, then that is an interesting data point to consider when choosing suppliers.

This is constructive criticism. The people that are the most upset about the situation will just change brands rather than post complaints. And the people that are upset are some of your larger customers that have a big installed base to worry about.

George

edit for tone and clarity

I am saying you make ONE, just ONE, Routerboard that has a modern processor and fits the 532 form factor. There is a decent market for this product that will be there for years and years to come.

Could you give some more details on this product as you invision it? Do they just create the "432" and let it be? When the 400 series is advanced to an updated version of the processor in the future (faster, new chipset features, etc), do they update this old form factor version as well? How long do they support this hybrid product? If they don't support this version forever, how does Mikrotik get the word out that this is a temporary product only that won't be supported forever? Or are you suggesting Mikrotik support both form factors from this point forward?

Scott

I would be very happy with a 432 LTS that progressed through multiple revisions like the 532 did. If MT decides that there is enough of a market, maybe we would see a 732 LTS or something later on and it would start off with a four year commit too.

I suspect the bigger issue is going to be moving to ExpressCard slots in the future when better radio cards are available but that may just be a case of 'tough, move to the new form factor to get them'. That attitude would be completely understandable as its an upgrade rather than a replacement.

From my perspective, support and availability on an LTS RB version doesn't have to last forever as long as its defined up front. Three years, four years would be great. If you are buying industrial computing boards, they seem to have about a four year availability window for example.

If you commit to an LTS version of Linux, you know when its going to be EoL'd. Same goes for Windows for that matter, they just did a final EoL on Windows 3.11 apparently.

From a customer-facing business perspective, this is all about 'stickiness'. I know John Tully gets it, but I'm not sure this decision to abandon the installed base was thought through properly. If we have to change out everything up the tower to get the new form factor, the only 'stickiness' left driving that purchasing decision is RouterOS itself. And RouterOS X86 can be run on alternative hardware so we don't get burned again. And once you are on X86, there are software choices other than RouterOS so MT loses that 'stickiness' even more...

In fact, its actually worse than that as there ARE other hardware solutions that will fit existing enclosures, so there is a net disencentive to stay with Routerboards. Scary...

Ya gotta look after the base... Never make it an easy decision for a customer to change suppliers!!

George

A couple more points. First is that MT is very decent in providing a relatively unfiltered environment to air these concerns. Some of their competitors would not.

My latest Routerboard shipment literally just came off the UPS truck and is sitting by the door. The current 433 is an excellent workhorse and we're very pleased with all the current products.

Our only issue is how to replace all the older existing gear in the air without having to change the world, so consider this a new product request rather than a complaint against the current product line.

The single biggest problem is an unscheduled repair to old-form-factor equipment. A thunderstorm causes problems to an otherwise happy 532 for instance...

It takes much longer to fix today and costs a lot more time and money. At least with upgrades, you can schedule them and allow time for changing out enclosures.

Please note that the original poster and other contributers to this and other form-factor threads tend to have been MT customers for a long, long time. I think I'm member 10 on this forum, and half those ahead of me are from MT!

George