I recently had to design a solution to power a arm based system like arduino with an external power supply but, as this system have to stay up even if current is cut, I also had to backup it with a battery. This post is describing the design for a such circuit.
Internet is providing more and more bandwidth to transfer files and files are bigger and bigger. One thing is not changing, it is latency as distance still the same over the time. When a protocol requires acknowledgment between blocks of transfer this latency is limiting the throughput like explained in this post.
The throughput is really different depending on the protocol in use to transfer the file. As I did not found something giving a lot of data to compare the existing protocol, i’ll try to get figures myself and detail here.
My last post was about a demonstration of what you can do in 5 minutes with MQTT. This video was part of a global MQTT presentation I’ve done for Clermont’ech APIHours.
Here is a full version of this talk in English. For French reader, the video of the talk in French will come soon as soon as Clermont’ech team will finalize it.
Enjoy !
Slide are here : http://slides.com/disk91/mqtt#/
To prepare a presentation of MQTT for tomorrow, I just finished a MQTT / Mosquitto demonstration in a 4 minutes video. Enjoy !
Synology is providing a Glacier backup tool which seems to be an easy way to backup data on the low cost amazon solution. I will not describe how to configure it a log of blogs are doing it and I do not like to be redundant. The question I did not find an easy answer was : will I be able to restore my backup even if my synology crashed and I would not be able to get one other…
Recently I killed on of SigFox Td1208 device by misprogramming it, destroying the bootloader. In a such situation serial port is not anymore accessible for programming. The use of JTAG port is a necessity. On the TD1208, jtag has been repaced by SWD which is a 2 wires port doing the same thing.
To use this port, you generally need a specific interface. These interfaces like Jlink are expensive (> 300€), some are less like BusPirate (30€). By the way, if you do not have it, you will have to order and wait for it. That’s why I’ve taken the choice of using my BeagleBone black device GPIO to pilot these signal. This code can easily be port to RaspberryPI…
There was something I missed on the IBM strategy when they sold x86 branch to Lenovo. Since I read some articles about OpenPower and Google home made first power8 server, this strategy is making more sense.
Actually on the server arena, we have the big one : Intel and front of it we had ARM, but if ARM is a good low consumption solution, working really well on mobiles/tablet and objects, it is far away to support heavy I/O workload and is not in a direct competition with Intel until couple of years. On the other corner we had the Unix proprietary systems and one of the last of them, IBM is considering by Gartner as a technology to be replaced after the next 2-5 years. Solution that you never see in any cloud. The future was clear : x86 has to dominate the next decade.
In this context, understanding the IBM strategy to sell x86 servers to Lenovo was not really clear and looked like a kind of hara-kiri of the IBM hardware branch. From this last announced of Power8 and OpenPower, allowing other vendor to use Power IP to design its own hardware, the understanding of the strategy sounds differently. By selling x86 to Lenovo, IBM becomes a real competitor of Intel instead of being a large client of Intel expecting good chip price for its servers. Now IBM can (and have to) look for Intel market share in the datacenter and can (have to) enter in a direct fight.
This announce, only coming from IBM, would have been heard as the last tentative from a dieing entity trying to open its technology to make it survive. But coupled with the presentation by Google of its first prototype of Power8 server, it gives a real credibility of this direction in a period where large Internet companies (Facebook, Google, Amazon….) are looking to build their own hardware supporting openHardware initiatives.
In my point of view this announce is a really good thing, mostly if you consider (like me), that Intel based server are good and not expensive but not reliable enough and not really vertically scalable. Reliability is thing OP guy’s and business are really considering as important. The first coming question is about the software platform on top of it ? I do not consider AIX to be an answer, but Linux is. In this direction all the existing Linux ecosystem has to be recompiled / adapted for this new platform to take most advantage of it ; seeing Google going that way is a good news as they are an important actor of the datacenter Linux ecosystem.
Let’s wait, now, month pass to confirm if this was a bankable strategy for IBM or not.
Pir sensor
In a previous blog post I described how I built a PIR sensor connected to Internet, based on the use of a Raspberry Pi device. As you can read in this previous post, the solution have a certain number of complexities like requiring a power supply, a 3G modem …
As it was originally planed to be done, here is the design of the same product, based on a SigFox device. This is really interesting to demonstrate the differences and the advantages of this technology. Let’s review how to built it !