On 27/10/2014 00:18, Carsten Haitzler (The Rasterman) wrote: > On Sun, 26 Oct 2014 18:13:14 +0000 "Saxena, Sunil" <[email protected]> > said: > >> Agreed that IoT will require us to add to existing Tizen Common for >> discovering, communicating, security, services of these devices. >> >> I wanted to highlight a different point below. >> >> When we defined TIzen Common, it was defined to be constrained as a base for >> devices that required a minimum of 256 Mbytes and more likely 512Mbyte >> device, which are IVI, Mobile, Phone, TV, Camera etc. This is where all the >> work on Tizen.org is focused on and let’s call it Tizen Common Full for >> discussion sake. >> >> Going forward you may want to create another Tizen Common Micro that may be >> constrained for devices that require 64 Mbytes or less which may be for >> devices like thermostats, wearables etc with almost no display. This profile >> would contain minimum core packages and functionality and would be subset of >> Tizen Common Full. >> >> Of course, you can go down further to 4 Mbyte or less devices and possibly >> create Tizen Common Nano – terms for discussion sake.
If we look at the estimates - e.g. Cisco predicting 25 billion IoT devices by 2015, 50 billion by 2020 - well there's a question as to whether those estimates are overly optimistic. Leaving that aside, as mentioned previously, the IoT is a marketeer's dream. It means everything, and thus nothing, but it is getting people enthused. :) More so than "pervasive devices" did or "machine to machine did" or "yeah we put TCP/IP stacks in embedded devices" did. It's important to remember there's nothing new here. People have been doing this for at least 20 years. _If_ these billions of devices appear then I am quite certain that the very vast majority are going to be small form factor, very low power (battery or energy harvesting) sensor devices. To me these devices are best captured as "embedded, connected devices, operating within a utiility infrastructure". Utility is important. These are not smartphones, are extremely unlikely to have more then a button and an LED as a UI and you're not going to be able to get to them to turn them off and on again. As an example, the kind of devices we're developing have 128Kb (Kilo, not Mega) of non-volatile storage available, half of which has to be reserved so we can update the devices in the field without potentially bricking them. So the very vast majority of IoT devices will not support Linux, or even ucLinux, and Tizen has no place on them. However I do believe all, or most, of these devices are going to require edge gateways to map between the mesh network and the Internet. (IPv6 may solve that problem transparently but I doubt it. Particularly as I have yet to find anybody with a sensible network infrastrucure answer to the power budget question. I believe we need edge gateways for now). Embedded Linux devices and Tizen devices do, I think, have a place here. The type of devices we are developing are both with and without screens, and we've been looking at Embedded Linux, WinCE and Android, so I see at least two two profiles there, perhaps similar to the Home Controller I've seen mentioned. Cheers, Alex _______________________________________________ General mailing list [email protected] https://lists.tizen.org/listinfo/general
