Rude awakening

My credit card was charged when I registered, according to Microsoft it would be reverse at the end of the month. Will monitor that to ensure the reversal is done.

After playing with the board I realized that the  free subscription will be used up really fast  if we are not careful. If this were to go into production you will have a few hundreds or worst  a few thousands of these boards deployed as one of my customer pointed out.

So my testing would Really determine whether it's feasible to go with this board or some other SBC/microcontroller that doesn't use Microsoft Products( restricted for now).  And implement some other security like WoTT.

Just my personal observations  after testing..........

Your observations provide insight into an incremental costs that would come into play post development for a production solution. The total cost of ownership (TCO) for such a solution would need to include the data overhead costs.

Is the Azure solution restricted to just the MS portal or does the technology permit other opportunities? I ask because I don't have any insight. I elected to not participate in the offering. In the end Canada was not supported in the trial.

I make it a point to closely examine TCO for a solution I have found sometimes the sticker price for implementation is attractive but the ongoing costs can be crippling. Especially when the pricing is per unit deployment.

I recall implementing a vendor solution that included 25K licenses. No one asked about the costs posts 25K. After being in production for two years the initial license block was consumed and the size of addition license blocks was smaller with the costs incrementally higher. What could you do when you are locked into the solution!

Something designers should be aware of. IoT solutions for ten development units verses 10K production units can exhaust a budget real fast.

Sean

Your observations provide insight into an incremental costs that would come into play post development for a production solution. The total cost of ownership (TCO) for such a solution would need to include the data overhead costs.

Is the Azure solution restricted to just the MS portal or does the technology permit other opportunities? I ask because I don't have any insight. I elected to not participate in the offering. In the end Canada was not supported in the trial.

I make it a point to closely examine TCO for a solution I have found sometimes the sticker price for implementation is attractive but the ongoing costs can be crippling. Especially when the pricing is per unit deployment.

I recall implementing a vendor solution that included 25K licenses. No one asked about the costs posts 25K. After being in production for two years the initial license block was consumed and the size of addition license blocks was smaller with the costs incrementally higher. What could you do when you are locked into the solution!

Something designers should be aware of. IoT solutions for ten development units verses 10K production units can exhaust a budget real fast.

Sean

Hi Sean,

All the cloud providers have billing that sometimes takes a while to get the head around, not because they're intentionally trying to get the casual user testing their system to pay huge fees, but just simply because the casual user is using the exact same system as the people building large services.

They don't mostly mind the casual user with their few hundred requests a month, because they want them to build up to hundreds of thousands of requests a month. My monthly AWS bill is tiny, I only use it to test a few features from time to time.

Once with AWS as an individual I had a huge bill ($200 for a month), but it was my fault - I was using IaaS (i.e. dedicating resources for myself), and I powered up a massive machine, and forgot to turn it off : ( In that respect, the IoT solutions from Amazon, Microsoft etc can be cheaper for the causal experimenter or design engineer, since it is PaaS which is not dedicating (much) computing resources until they are needed. Taken to the extreme, there's the possibility of consuming zero computing power until customers come along, using so-called FaaS (function-as-a-service). With that, provided you're not using other billable things to design the solution (e.g. large databases) then there is zero cost (or near-zero, since it depends on how the service provider wants to handle it) and you only incur cost that grows as usage of the service grows. AWS offer that, Azure will have a similar thing.

Using typical third party software IP in code can (sometimes) initially cost heaps more than an IoT solution from the SPs, depending on how specialist or desirable it is. They like older licensing schemes, e.g. a block of licenses, so there is a larger up-front cost.

The IoT solutions are more granular, at least the ones I've encountered, to encourage developers to test them out with zero or near-zero cost, and then ramp up.

I have limited experience developing systems outside a closed environment. Closed in the sense the systems were within a network the company owned and managed. As an ISP provider the network was large and 100K of devices. Not big to US standards but large in Canada.

In the true sense of IoT, the client (i.e. IoT device) could be in any network. Finding an IoT data collector that has a reach into the expanse of any everywhere network would be daunting. I can only imagine the claims of vendors to garner business. Proving the claims would be a task.

Providers of the service and costing models piques my interest. I have to confess, I can have an interest knowing the details but am relieved I won't be part of a team trying to deliver.

Sean