lauantai 15. huhtikuuta 2017

Spareparts

We are starting to build the new file server today. I'm still missing the intended CPU and a better cooler, but it turns out Olli the overclocker has a box of spareparts...


Photo credits (c) 2017 by Jari Arkko

10TB

I am in California for a quick 2 day visit this February, on IETF business. But I'm not so busy that I couldn't stop on the way from the airport to pick up some storage from Fry's :-)


Photo credits (c) 2017 by Jari Arkko

sunnuntai 3. tammikuuta 2016

Revenge of the Electronic Shower


If it has electronics, it can break. Like the my teenage son's shower unit...

Oras Ventura is a battery operated shower controller than has a display and drives the water controls via electronic motors. I have two of these in the house, and they generally have given very little grief in their soon 10 years of existence.

The units display temperature, and strength of shower can be controlled via up/down buttons, as well as choosing hand/top shower. Not a big deal as such, and other than looking great they don't add that much functionality. One potentially useful function is that there's an automatic 5-minute cut-off for the shower, if it is left accidentally on. Possibly useful for people like me who worry if they left shower/oven/iron on when they leave the house :-)

The downsides are that this brings more parts that could break, and the potential for battery changes. Also, the Venture seems like a first-generation product. The response to inputs is sluggish. And you can't load any shower apps :-) And there's no touch screen or WiFi. WTF?

But the four D20 batteries last a long time, in the order of five years even with daily use. However, now one of the units is acting up. Not turning on the display, making a constant noise, and not responding to inputs. It first displayed E10 (per unreliable teenager reports) and now displays nothing, even with good batteries.

Revenge of the Things?

tiistai 8. syyskuuta 2015

More Health Gadgets - Withings BPM



Yay, another gadget arrived today! For a long time, I've been a Withings weight scale user, and have liked the device a lot. But now I got a blood pressure meter from Withings. This is a review of the device, based on my initial experiences.

The Good

First things first, the design is quite cool, the hardware is sturdy, the finish is perfect. The overall idea is great, your measurements sent to the Internet over Bluetooth. The user interface for looking at the measurements is easy and looks great. Furthermore, Withings allows you to download their data easily, as well as upload other measurements to their database. This makes moving your data to your own storage or other providers easier; it avoids the customer lock problem.

And, most importantly, two days have passed with this device and I have made more measurements than I made with the previous equipment all summer. I need to take measurements and have an idea later what numbers I have seen. Having to open a computer to take notes is make-work.

History graphs show developments over longer period of time. Like in this case where (sadly) my weight keeps increasing:



The Bad

There are some complaints, as well, however. A practical issue was that it took half an hour for me to get the device to work with my iPad. Android phones, iPhones, and iPads are all listed as supported devices, and the connection instructions are simple - pair the Bluetooth device, then push the start button and wait instructions from the application.

Except that in my case the application did not react in any way. Nothing. The device was paired, but not recognised by the application. The instruction booklet said nothing useful about this case.

After some googling it turned out that there's a bug that prevents the measurements to start, if the application is already running. But exiting it didn't help; the application did not start when I pressed the button the device.

Some more googling revealed that while iPads are supported, there is no iPad application. One has to download an iPhone application, and only that application supports running the measurements, even if the iPad application supports showing all measurements. And that was the one that I had used for the weight scale as well.

The software installation initiated by Bluetooth pairing was also cumbersome. The pairing takes you automatically to Appstore for loading the application, but to a place that finds no applications and offers just wrong choices for what to do next (like redeeming a voucher). The instruction booklet says that you should either install the application or update it, but nothing about the case where the application is already installed and most recent version. And that is of course the likely situation if you already own a previous Withings product.

The silly thing is that all this could have been easily avoided if the pairing process would have downloaded the right application or if the instruction booklet would have included separate instructions for iPad users.

The other annoying this is that the meter has a "Start" button. But that button doesn't actually start the measurement -- it just starts the application, and then you have to press the real Start button in the application.

Internet of Things Architecture Thoughts

This is just one of the many examples where we buy things that are bonded to our smartphones. You cannot operate them without your phone. Granted, you are not likely to stop carrying your phone around any time soon. And the ability to see the measurements on high-quality user interface is very helpful.

That being said, I cannot help wonder if there's a better way to do this. In particular, there are many devices that are not designed to be used on a person; they could not rely on a smartphone in any case. And even for the blood pressure monitor, I would just like to take a measurement, and not have to press buttons on my iPad.

What the smartphones are useful for is the initial configuration, and the user interface when you need it. So perhaps instead of being a mandatory part, a forced hop on the path of packets, maybe the smartphone could be the way to configure these devices to use the relevant mobile network or wireless LAN, and then use them for the user interface when they are present? And when they are not present, the devices could just work, and connect to the Internet directly. Like this:



This would also be a great way to configure devices that you do not carry around, e.g., home appliances. Configure once, let them do their thing on the Internet on their own.

Conclusion

This is an overall nice product, well produced and easy to use once you get going. I keep wondering if there would be a way to make it similar to my weight scale; no iPads or iPhones or anything needed. Just the Wireless LAN connection.

You can buy the BPM from the Withings webshop (but be careful to choose the right webshop, there are different ones for different continents).

Photos (c) 2015 by Jari Arkko.

lauantai 18. heinäkuuta 2015

Fast Cars and Networks



How fast can you make a 23-year old Volvo 740 go? Turns out that you can make it go pretty fast... at least 200 mbit/s, maybe more.

I recently got interested in maintaining my old car. Well, it was either maintain it or take it to the scrapper. I chose maintenance, and decided to make two important actions: first, get rid of the rust and repaint the car. And second, add high-speed Internet. With IPv6, of course.

I acquired an LTE Advanced subscription from my local operator (DNA) and a router that can support fast connections. LTE Advanced is a new variant of the 4G LTE standards, capable of speeds up to 300 or 450 mbit/s, using techniques such as carrier aggregation.

To be clear, these are some very fast speeds. My current ADSL connection at home is 18 mbit/s, so there is potential for 10-20x improvements. But the deployment of this technology is just beginning, so it is available only in select areas, and of course, carrier aggregation works best when there's enough spare capacity in the network and not too many other users. I do not yet have a lot of experience about this, but in most places I get 50-80 mbit/s easily, and saw also speeds approaching 200 mbit/s. Cool!

Someone asked what I use the network for. Well, to be honest I don't know. Once the technology is there, ideas for applications can come. What Internet function would you like to see in a car?




Details: I am using Huawei E5186 as a router, and have generally good experiences with this device. It is easy to setup and both IPv4 and IPv6 just work. Some configuration is possible, but isn't strictly speaking needed - the system works out of the box.

For IPv6, the DNA service provides IPv6 for all customers, and as long as the router supports this, devices attached to the router get it too.

On the network side, I believe DNA uses Ericsson network equipment.

For fast connections, the internal network in the car needs to be fast too; I'm using 802.11ac, so the router and devices need to support that.

All this equipment is of course mostly made for non-car use. The router draws 2A of 12V. I'm a bit reluctant to use the car's electricity directly, as the voltage tends to vary quite a bit. Unfortunately, voltage regulators for 12V to 12V weren't in the stock for the local electronics store, so I went for an inverter to 220V and then back to 12V. Not pretty, and uses too much electricity, but it is a temporary solution.



Photos (c) 2015 by Jari Arkko.

maanantai 29. syyskuuta 2014

Making Networking Easier for Users


One of my pet peeves is how hard it is to connect to various networks. I am on a trip today, and my first stop was at the Helsinki-Vantaa airport. A modern airport that prides itself on efficiency and good infrastructure. Their Wi-Fi has worked very well for me. I came to the airport, launched my browser, and confirmed my connection to the network with one click. The network is fast, has no time limits, and is available everywhere within premises. (It would probably work even in the newly opened airport sauna, if my tablet could take the heat and humidity.)

And, having travelled a lot in recent years, this network really is one of the best airport Wi-Fi networks in the world. So all is well, right?

No.

Many people, me included, have data subscriptions on their devices. This allows cellular networks to be used when no Wi-Fi is available. So I can get Skype messages, send photos, and do anything I like everywhere, or at least within my home country. There is only one place in Finland where my smartphone stops working: the Helsinki-Vantaa airport. The device sees the Wi-Fi, connects, and stops, waiting for the user to make that one click.

Of course, if I need to use the Internet I will notice all this and make that one click. However, I believe we are increasingly in an always-connected world, and information flows when it needs to. I do not want my devices to stop, I want them to keep on working. Even that one click is too much. And it is unclear what value that click brings anyone, me or the airport.

But the Helsinki-Vantaa network is one of the best. My next stop was a large hub airport in northern Europe. I checked if there was local Wi-Fi, and there was. My browser told me that I could use either the free or the paid network. I decided to try the free one first. A screen comes up that requires me to register. But the amount of information that they ask is... at least tiresome, if not privacy insensitive. The e-mail address and password I can understand. Maybe the phone number, because some airport networks send a password over on SMS. But I also had to fill in birth date and country. Why?


But this was just the start of my failed attempt to use the network. I got an error message that my chosen e-mail or password was invalid - without any explanation of why they were invalid. I guessed that perhaps I had entered an e-mail address that I had used on my previous visit, and entered another one. Not wanting to fill all information for the second time, I used the browser back button to get to the registration page. Big mistake! After changing my information and pressing register, I got an error.  I should have started again from the beginning.


I gave up and decided to try the paid network instead. I clicked on the payment page, but had to stop for a few minutes to walk to my gate. At the gate, I entered the required information. Including my mother's maiden name. Why? Having fulfilled all information, I later realised that my walk had made it impossible to complete the transaction, whether or not I was revealing all my personal information. I got a "your session is no longer valid" error message.


I could go on, but you get the point. The problems with these networks are very wide spread, and in my opinion, unnecessary. Earlier this summer I was waiting for my delayed flight at a major North American airport, and had trouble getting more than a few packets exchanged via their local network. Not enough to download any e-mails, for instance. Later a friend posted a picture of what his wireless network analyser showed at that airport: all the airport's access points were on the same three channels, making the radio environment in practice unusable.

Couple of years ago we had a large meeting in a hotel in Paris, and our network engineers ended up redesigning the hotel network to work much better, in a situation that was similar to what we saw at the above airport.



You might argue that I am talking about a first world problem. Travel. Airports. Hotels. But take the airport Wi-Fi situation as just one example of a more general problem. We all use networks, in developed and developing world. And many of those networks are often in shopping malls, coffee shops, and bus stations. We need to improve their usability.

The smartphone revolution that we have seen last couple of years was initiated largely thanks to the new, easier to use touchscreen phones and operating systems. Perhaps we need a similar revolution in wireless networking. And I think we have had a lot of the technology around but now it is time to put it into use. Examples of this include Hotspot 2.0, 802.11u, EAP, and various network connectivity checkers in our devices. (I worked on EAP at the IETF over ten years ago, but it is only now that it is getting more widespread adoption. A long time!)

And this problem is not limited to Wi-Fi. Again that is just an example. I also have a lot of experience of trying to acquire SIM cards in various places. I've had great experience in some countries and with some operators. Others struggle in providing a service. Some operators sell SIM cards and the needed data allowance package in different shops. Why? And the top-up process can be difficult. An operator in Europe requires a local phone number for the process to complete. Not particularly useful for tourists, is it? An Asia-Pacific operator uses a web-page process that only works on some browsers, not all. Many countries require proof of identity before you can use a network. Perhaps I can understand that, but it does not stop there. I was once asked for a formal proof of my address within the country. How can I produce such a proof, when staying at a hotel?

Here are my guidelines for providing an effective Internet access product:
  • If you want to provide an open network, make it truly open.
  • If your network is for subscribers, employ an automatic authentication mechanism that does not require the user to perform a web login.
  • If you charge for the use of the network, make it easy to provide the credit card information
  • If you sell cellular access, make the right size SIM cards available to everyone easily.  Do not force users to go through extra steps to "enable" the SIM card. You sold it, it can be enabled automatically.
  • Any web-based purchase, top-up, or login process needs to be tested with all browsers and in different conditions.
  • Do not collect information that you do not need.
  • Radio networks need to be tested and configured properly.

In short, people want your service. Make it easy to buy and use it!

Jari Arkko

Photo credits (c) Jari Arkko and Joe Abley

keskiviikko 27. kesäkuuta 2012

IP for Smart Objects

At Ericsson's Lab web site, Jan Höller and myself have published an article. This article deals with the general architecture for the "Internet of Things", and how we must break the current application-specific silos, and move to a model that employs general-purpose network access technology, IP & IPv6, and web technology. The link to the article is here.