Wednesday, January 27, 2016

I just realized I hate computers

Not really but there are aspects of computers that I no longer like or even appreciate and while I think I have evolved over time there are elements of the industry that have not learned a damn thing.

In the early 1980's I reversed engineered Dancing Demon in order to figure out how they accomplished the speed that was simply not there in BASIC.


This program used BASIC as a sort of bootloader for raw byte code which is where the performance came from. Later I would spend a lot of time in CPU and board design, copy protection, simulators, device drivers, TSR, operating systems, BIOS and maybe some other things I've long since forgotten.

  • Transec systems I used an ICE machine to debug copy protection and bypass it. Later I wrote low level DOS extensions to the filesystem.
  • IBM MSD had used for a number of systems including a highspeed simulation of a cluster of single board computers, BIOS for it's Artic Coprocessor; later I moved to the OS/2 team and worked on Video Drivers, AIX for PC, and WorkPlace OS Presentation Manager.
  • At Core International I implemented the disk and tape device drivers connecting the company's hardware to OS/2.
  • NaBanco I implemented a terminal messaging network as a DOS TSR.
What I'm saying is that over the years I have been writing code at a low enough level to understand the hardware and the operating systems. AND I LOVED IT. Tinkering at such a low level was always a thrill. It was especially fun when debugging a problem required an ICE machine, logic analyser or oscilloscope.

Fast forward 35 years; rotate in my chair to my workbench and I see a Dell C6100. It's quiet now and it might remain that way for the time being and longer. Between the IPMI, BMC and dated BIOS I've realized a few things.

These clunky machines are not the future. The cases weigh too much. The power supplies are too big. The Fans are mechanical, noisy, and prone to failure.
And this thing(chip) costs $9.
While the Dell C6100 cost $300 that was not it's original retail price. But regardless of it's original source, likelihood of a MTBF failure, it still requires plenty of sweat equity to get it up and running. More than I'm willing to put into it. I have moved beyond my interest in all things hardware. It also means that, at least Dell, has not learned anything about it's own hardware. The BIOS while simple is complicated and not necessary.

I think I want to use my phone as my computer.

And something similar on the serverside. Maybe some stackable Chip.

Monday, January 18, 2016

The Tech Garage Economy

I was sharing shark bites with a friend of mine when I came across this article.


Lore Harp CEO makes the statement "impossible for companies to get started today. the costs involved are going to be prohibitive. when we got started out we could do it with $5000 and a lot of sleepless nights..."

It's funny to hear this because just a few months after the iPhone revolution people said the same thing. And it was said again after the Android and iPad launches. It's possible with every generation of new device if you get in early enough.

iTunes will break in one generation

By one generation I mean a human generation not hardware generation. Consider that Apple's iTunes started as an "individual" user application. One song for one user. Shortly after that people could share an AppleID so that they could share content. In fact that's how my wife and I shared for a while. A few months ago Apple introduced "Family Sharing".  It's not hard to understand why. Our kids were getting hand me downs and we did not really want them to have access to the exact same software. But now as our kids get older they are going to want to have their own families. Sure, family sharing works in college and maybe during those single years but when our children have children of their own it's going to get difficult. It's even going to get worse when you consider divorce. "Who gets the iTunes account".

Observations - VPS hosting and the aftermarket

I just purchased a Dell C6100 for $300 including shipping and as I was sitting around this morning thinking about how I was going to handle the IPMI password problem I started to ask myself about why there were no drives included in the purchase. DUH! The machines were likely part of an end of life purge at some VPS. According to the seller they have sold hundreds of machines so I wondered where they sourced the machines that they could sell them for so cheap; considering that they had to have some markup to make money.
"hundreds" of machines probably means that they sourced the machines at Peer1 or maybe Rackspace. They would not be sourcing from Google because they design their own machines.
As I was considering my purchase I recall a conversation with a coworker. I was against the purchase for production because we have a lightsout operation and frequent hardware failures would only increase our costs. If "we" were willing to give up our free time to repair used systems then we might be better of directing that time a revenue opportunities instead.

Well, I'm looking forward to my new arrival. 4 systems in a 2U rackmount with rails and 96GB ram. In all I should be able to build out my system fairly easily although I'm just a little concerned about the noise from the fans. Given the price I should be able to cut the expense of my Google Compute and Digital Ocean instances. In fact the savings could let me install a second network in the house for a little redundancy and speed.

In all, just on the cost savings alone (a) I could host myself (b) buy enough backup hardware for those eventual failures.

Saturday, January 16, 2016

What happens to denied rebate claims

You go to the store, you purchase a thing, you download and print the rebate form, you cut out the UPC code, fill out the form, copy the receipt, package it all up and send it to the rebate processor or manufacturer.

Sometimes the manufacturer uses a rebate clearing house that keeps tabs on the process and offers feedback. Sometimes you simply have to wait for the check or the dreaded decline letter.

But what happens when you are declined? On the one-hand the reason you were being offered the rebate was so that the company can capture purchase and personal information about you. And you get cash in return for your personal information. It's a contract stupid.

But what happens when you're declined?

Does the company still claim and get access to your personal information? Does the 3rd party processor get to keep and resell your information? Or, once there is a decline, the processor or company quietly deletes your information after sending you an decline letter? And what sort of arbitration is there?

My advice, forget the rebates and forget that model. Chances are better than even that the seller, or "the house" if you're in a casino, is making bets; and given that they know more than you... chances are it's a loser.

  • forget the rebates
  • ignore products with rebates
  • know your products and model numbers
  • comparison shop
And there was some advice I received once. Check the model numbers on google and if the results are biased to one seller or even some particular promotion then stay away from it. It's likely bait; meaning inferior features or components.

Then again, unless you're an early adopter stay away from full retail.

Amazon.com rebates

Then there is the rebates. As I mentioned in my last post that Amazon computed the rebate price even though the "purchase by" period has passed... Now here is some more Amazon magic.

As I was still looking for a good barebones computer for some clustering I wanted to experiment with I found another MSI device. This one appears to be $32 after rebate(s). That seems to be a great price if you ask me!


So when I looked at the rebate links I was not surprised to see that while the dates were ok... there were 2 rebates in the popup. So clearly someone made a mistake. Whether it was Amazon or MSI it's tough to tell but he rebates are identical and I seriously doubt that MSI is going to accept two rebates for the same purchase.

While this was not the compete terms and conditions (there was some small print at the bottom) This section looked familiar to many other rebates I've seen. The good news is that it's not limited to one per household and not even one per purchase.

Well, I was wrong again. When I read the fine print I found:
Limit (1) rebate per qualifying rebate offer, per person, billing address, company, household and receipt/invoice/packing slip during the eligibility period, except were prohibited by law.
I don't think this was intentionally misleading and in some cases rebates can make the difference. Keep in mind that rebates are the company's attempt to capture analytics that would otherwise be limited to the seller. While that makes sense on one device it would drive the margins down too far and encourage resale competition on eBay etc.

Finally, the UPC and the application need to be postmarked no later than 15 days from purchase. Since I previously noted that in some cases there was a 20 to 45 day delay it might be impossible to collect.

amazon.com delivery dates

Some time ago I complained about an order that I had placed with Amazon.  I was under the impression that it was going to arrive in a few days when it took a few weeks. I was not quite sure what happened and what I had missed but it happened again. And this time I caught it.

Here is the item in standard Amazon Search results

Here I am in my shopping cart

And then the final review

Finally, I logged into Amazon and reviewed my orders from the account orders screen and ...

While this order was for an MSI barebones computer, I had gone through the same process a few minutes ago for a different brand but in that case the FINAL arrival date was between Feb 15 and March 15. Needless to say I cancelled that order too.

One other complaint. May times there is a mail-in rebate program associated with a product offering, however, when the rebate has a "purchase by" deadline the item is not updated to reflect that the rebate period has passed. In fact one of the PCs I was going to purchase had a purchase 1/4/2016 deadline and it would have saved about 30%. (the other limits of 1 per household still exist).

Friday, January 15, 2016

rancher, docker and coreos requirements

I read some place that all I needed was 1GB or RAM. That seems pretty simple and if you look at the various systems from digital ocean and AWS then 1GB is still pretty simple, After I added sysdig my system ran for a day and then froze. A reboot failed too. ultimately dmesg indicated that there was no disk space remaining and until I determined that it was /var/lib/docker and /var/lib/journal it took a while to get the system back in service.

I would think that by now "people" (docker, coreos, rancher) know what the ideal server configuration is and not recommending 1GB is a good starting point.

UPDATE: I'm getting some flack from the owner of the rancher github account for reporting this as an issue. I've tried to explain that this is related to the VM configuration but I'm starting to realize something else... (a) when I deploy bare metal servers I usually allocate some multiple of 500GB of storage depending on the price of drives and RAID. (b) when I deploy an virtual host on a VPS I usually start at 20GB per instance. By extension that means that each docker instance should have 20GB. Given that docker has a host requirement that means I need to allocate more storage on the host side of the equation.

Wednesday, January 13, 2016

Flow Based Programming (FBP)

I have been a fan of FBP since I watched the first kickstarter video for noflo. Forget the metrics associated with CSV imports I saw a much bigger solution for a very real problem. My only objection was that noflo was implemented in javascript/coffeescript and while not unfamiliar was not going to take me where I needed to go which was writing complex server applications in addition to UX. I was finally able to implement a Flow based framework in golang which is now the basis for a number of applications including a reporting DSL, payment gateway transactions, and a few webapps.

While I was doing some reading for a client I found myself reviewing Meteor thinking that it might be a good framework for them... which then triggered my noflow memories. I went back to noflo and flowhub only to discover that they are not making much progress. The changelog suggests that they are in maintenance mode. I'm pretty sure that while FlowHub was supposed to be a PAAS (similar to Meteor's Galaxy) it turned into some other sort of platform which looked more like shopify than a hosted FBP. Furthermore the noflow commit log is also pretty slow.

Flow Based Programming rocks! In my experiments I have been able to implement the "steps" much the way someone might implement regular methods, however, having a framework to stitch it all together means that error trapping, metrics, reuse can be global.

Friday, January 8, 2016

Travel Routers Part 12 - Asus WL-330NUL

I was hoping that this device was going to be the defacto winner. It is clearly the smallest and lighest but it might also have the fewest features.

Asus WL-330NUL
Battery - none
Ports - Male USB for powering the device, there is also some storage for installing Windows drivers, RJ45 ethernet port.
Modes - Access Point, bridge
GUI - web admin
Default SSID - WL-330NUL-private-XXXX
Default password - (supervisor code etched on the device)
Default IP Address - router.asus.com
Default uid/pwd - (supervisor code)
Manual: link

After a brief use here is what I've experienced:

  • slow throughput; while the ping was ok the download speed leveled at 14mbps and should have been higher. Considering it's side it might have been related to the amount of RF in this location.
  • no VPN support
  • There is a guest user
  • I could change the default password and username (the supervisor code is unreliable over time)
  • the boot time was very slow. It took nearly 2 minutes to boot
and lastly the packaging was unusually large for the size of the device.

Thursday, January 7, 2016

CoreOS - discovery etcd io

The first step in deploying my next cluster is building a bootstrap server. This bootstrap server needs to host a number of small static services that are used by the nodes in the cluster(s) and in the worker pool. Examples of these services include, NTP, DNS, PXE/TFTP as traditional *nix services but then discovery.etcd.io as needed by etcd in order to discovery cluster membership.

coreos:  etcd2:    # generate a new token for each unique cluster from https://discovery.etcd.io/new:    discovery: https://discovery.etcd.io/<discovery_token>    # multi-region deployments, multi-cloud deployments, and Droplets without    # private networking need to use $public_ipv4:    advertise-client-urls: http://$private_ipv4:2379,http://$private_ipv4:4001    initial-advertise-peer-urls: http://$private_ipv4:2380    # listen on the official ports 2379, 2380 and one legacy port 4001:    listen-client-urls: http://0.0.0.0:2379,http://0.0.0.0:4001    listen-peer-urls: http://$private_ipv4:2380
** Sorry, clearly glogger did not paste the code properly.

The CoreOS team developed and deployed a public version of the discovery tool and then made the code available. Unfortunately the tool itself needs to be deployed in a cluster of etcd servers. And so there are two conflicts... (a) whether or not to use the public instance. (b) whether to perform the discovery manually.

(a) the TTL means that the record and UUID should not live long enough for someone to trick your cluster in order to replace one of your nodes into believing (i) that is does not belong; and (ii) that the bad guy can replace it. I'm not an expert but I imagine that one could validate the cluster peers with some list of IP addresses from a 3rd party proxy; in my case the "retrieve droplets" API at Digital Ocean.

(b) before implementing a pared down version of the discovery service and the reason I think that a lite version is required; read this doc as it describes the discovery protocol and hints as to how easy it might be.

The challenge is the design...

Create a private discovery service inside your firewall... but it requires an etcd cluster. And that cluster depends on tokens... so either you have to hand stitch the token or use the public discover service... which depends on an etcd cluster that was already clustered... and follow the tokens and services recursively until an ops person installed the tokens manually.

Because the TTL is fairly low, some temporary persistence, and because it's not necessary to stay alive 24x7 it would make sense that the discovery service might be detached from the etcd service.

UPDATE:  I received a response to a G+ question from Brandon @ CoreOS. "The discovery.etcd.io service only lasts for the purpose to construct the cluster. Nothing else." So the defacto discovery service is probably safe.  ON THE OTHER HAND the CoreOS clearly documents:

Running Your Own Discovery Service
The public discovery service is just an etcd cluster made available to the public internet. Since the discovery service conducts and stores the result of the first leader election, it needs to be consistent. You wouldn't want two machines in the same cluster to think they were both the leader.

Since etcd is designed to this type of leader election, it was an obvious choice to use it for everyone's initial leader election. This means that it's easy to run your own etcd cluster for this purpose.

If you're interested in how discovery API works behind the scenes in etcd, read about etcd clustering.

It's sort of a miscommunication here. Furthermore, creating a discovery service requires an etcd cluster which in turn requires an etcd cluster. And while there is some documentation on bootstraping an etcd cluster it seems involved and should have been scripted.

reference: http://bit.ly/1OUpWDc

Wednesday, January 6, 2016

Travel Routers Part 11 - Conclusion

There are a slew of other brands that I did not investigate. They include:

  • Buffalo - very small and no battery
  • Netgear - had a silly flap that exposed some LED and pretended to be an antenna
  • TP-Link - mentioned although I have a micro bridge here somewhere
  • D-Link - more expensive than the bar I set
  • Various 3G/4G devices - required a SIM and a service contract
  • Various non-name brands
  • Asus WL-330NUL - on order
  • Edimax BR-6258nL - looks like the Asus
I stand on the reviews I provided for the brands I tested. If I had to put together my go-bag I'm not sure which one I'd put in it. At the moment the front runner might be the HooToo Titan because of it's battery capacity.

UPDATE: With the Asus WL-330NUL on it's way and the docs I've read so far it's starting to look like the right device. I'm guessing that it also supports VPN Client but that would be a guess. I'm also convinced that the Apple products are the worst as they require the airport utility to perform the device setup. 

Travel Routers Part 10 - Apple Express

Up until May 2015 I had been a serious Apple enthusiast. While I have not owned a Mac Pro I certainly longed for one and I have convinced my employer(s) to provide me with Macbook Pro. I personally have had Minis, Airs, iPads, iPhones, Airports, and maybe some things I've forgotten.

About 2 years ago I started to become disillusioned as I was experiencing some issues with the corporate VPN solution and my Airport firewall. It was a drawn out exercise of support calls, emails and replacement hardware. The problem never went away and so I started investigating alternatives. ChromeOS.

Back to the Airport. I have owned 1 Airport Extreme, 2 or 3 socket Airport Express and 2 modern Airport Express (the former being a better design and more recently being copied).

The single biggest issue with the airport brand of routers is that they require an Apple computer, iPhone or iPad to configure. The second biggest challenge is that the first generation of Express, which is the most mobile of the devices, seems to have hit end of life.

PS: no battery included.

Travel Routers Part 9 - HooToo Tripmate Titan

I was going to write a detailed review of this device, however, since it arrived I've noticed a number of similarities.

  • button
  • battery level
  • WiFi/power indicator
While the original HooToo tripmate does not have port covers the RavPower does. Interestingly the HooToo Titan does too. I'm not sure what the use-case is for the port covers because It's not like I'll be using submerged although at 10,000mAH or even 6000mAH it would not take much to generate heat or start a fire. The HooToo Titan is proportionally bigger than the Tripmate original.

The thing that bothers me most is that the firmware versions of all 3 devices are the same even though the HooToo Tripmate is distinctly different and the HooToo Titan is almost identical to the RavPower. Let me state, again, that the version numbers are the same, the user manuals are nearly identical, and the support phone numbers are the same.

PS: I did receive a response to my support request at HooToo. The person I communicated with is clearly not an English speaker and the information I received was clearly non-responsive.

Hardware Scale

I need to get some hard facts but one thing that is bothering me about hardware costs is that, all other components being equal, 4x16GB systems costs less than 1x64GB system. I ask the question because I want to build a small cluster of CoreOS systems but realizing that in the long run it might be advantageous to have bare metal and local systems I started thinking about costs.

Travel Routers Part 8 - TP-Link

Since I discovered that HooToo and RavPower seem to be the same company I decided to return them and buy another brand.  Just on principle alone. I decided on TP-Link as I have already done some research on them previously and that Google is using them to build/manufacture one of the OnHub versions.

Well... even though amazon.com seems to be selling the crap out of TP-link and that there are so many many models including 3G and 4G... their own website seems to be omit these devices. At the very least I would have expected to see support results from the search.

So I'm calling TP-link a dead end.

Tuesday, January 5, 2016

Travel Routers Part 7 - HooToo Tripmate Original

The first of the 2 HooToo travel routers arrived today and at first glance it looks similar to the RavPower. The front has a button with a quick battery level indicator and a WiFi/active indicator.

HooToo Tripmate Original
Battery - 6000mAH (quick press the power button to see power level) While the documentation calls it an "external battery" it seems to also power the router.
Ports - mini USB for charging the battery, USB port for charging other devices and flash drives, RJ45 ethernet port.
Modes - Access Point, bridge, file server
GUI - web admin, smartphone app
Default SSID - Tripmate-XXXX
Default password - 11111111
Default IP Address - 10.10.10.254
Default uid/pwd - admin/(no password)
Manual: link

** The only difference between the HooToo HT-TM01 is the SD Cardslot, the size of the device, the theme of the web interface. I originally thought that the web interface theme was different, however, it's not.  In fact the GUI is exactly the same as the RavPower. The one difference is that (a) the WiFi LED colors are inverted from the RavPower and several of the pages are returning server 501 errors. Looking at the download page I see that the version numbers are also identical.

After doing a little research I was able to determine that the support phone number is identical between the two companies(ravpower, hootoo).

UPDATE: Correction; the webui has a different theme. It turns out that my browser was caching too much from the original router. Clearly the admin tool did not set the TTL or cache aging correctly. Well, they are the same device but for the SD card. I'm going to assume that they are the same under the covers.

Monday, January 4, 2016

Travel Routers Part 6 - KissLink

The KissLink router is probably the easiest router reviewed so far but not without it's challenges.

The connection process is pretty simple, just touch your NFC device to the KissLink and you're just about there. If you have a static device like a laptop it's almost as easy to connect. There is no fileserver; but who cares really? The Android application is OK but far from rock solid. Depending on the connected WiFi network the app get's lost. Also updating the firmware requires that the "device" is the admin device before the messages make any sense.

  • router(requires ethernet) and repeater/bridge mode
  • Clone MAC but who cares
  • white and black list
  • no logging
  • what does the "zen" button do?
  • no VPN
  • too big to be a travel modem
  • uses USB power
  • no QoS
  • Since the documentation talks about "static" devices like laptops I'm not sure whether the Chromecast setup is going to work. 
  • hidden SSID
I'd like to say that the KissLink application is useful and unique but it's not. They tried to look like OnHub or vice versa but wins the close but no cigar. The more error paths I went down the more I realized that the basic documentation was incomplete or popup error messages were just wrong.

There is something to be said for being able to add devices on the fly based on NFC. Considering how often we travel vs how often we have no devices. It is also interesting as a way to let guests into my network and in bridge mode I can locate it in more accessible location.

Saturday, January 2, 2016

Travel Routers Part 5 - early pro/con

After doing some testing I have made some early observations.

ZyXEL - pro
  • small and light
  • several modes
  • WPS button which I have NEVER used
  • no battery - nothing to recharge or discharge
  • supports DDNS
  • firmware is current suggesting it's stable (2013)
  • some logging
ZyXEL - con
  • no battery therefore impossible to be mobile
  • many tasks require reboot and it's not clear from the menu
  • no users per se
  • no VPN
  • rebooting can require multiple reboots
The biggest hindrance to the ZyXEL seems to be speedtest.net and QoS. I tried a speedtest with QoS disabled and the performance varied widely. I suppose there could have been some spectrum competition with my neighbors but the results were consistent and so I think it's the device. When I enabled QoS the device took a dump. The GUI indicates that the QoS bandwidth is measured in kpbs, however, never really go there.

FileHub - pro
  • speedtest was normal and nearly full speed.
  • supported hidden SSID (great for hotels)
  • GUI auto detected and displayed a popup when I inserted a USD drive
  • into my second hour and the battery indicator is still on 3. It must be a curved scale so that 4 is 100% and 3 is not 75% (25% per indicator) so that 76% is not considered full. Correction, the documentation says 4 LEDs is 75-100%. I guess the factory change was just over 75%.
  • confirms in the DOC that Chromecast is supported
FileHub - con
  • there is an option to change the hostname. I suppose that the hostname must be an indicator to the upstream network device but it's not crystal clear. I'm not sure I want to identify myself either.
  • Changing the SSID caused the GUI to appear to hang which only meant that the SSID changed and I had to update the connection; not so obvious but not unusual
  • no file download feature but I can upload. Samba support would normally be a pro, however, since I can already upload but not download samba is just a curiosity. 
  • SECURITY - the guest user was ON by default. 
  • SECURITY - the samba service was ON by default.
  • disabling samba displayed a curious warning message, however, disabling the guest user did not.
  • no VPN
  • only the admin and guest user; which is not a strong 'con' but it is not immediately clear what a user can do
  • removing a mounted USB flash drive meant navigating to the "information" menu. This was completely counter intuitive.
  • not certain what happened when I tried my USB HDD. Since the drive was a MacOS filesystem it should not have been readable, however, an error would have been nice.
  • upgrading the firmware caused some features to be reversed...(check those security settings)
  • no logging at all. How am I to know if I'm under attack while traveling. The Router is the first point of attack and if penetrated could leak everything.
One thing I learned from the Sandisk Connect and Sandisk Media Drive is that one can conserve a lot of battery if you download the media instead of playing over the air. Granted the FileHub has a big battery but then things get complicated when the battery is empty and you try to use your laptop's battery to prime the device just to get a few files from it. On the otherhand the media is portable itself so if you are out of battery you can move the media to your laptop. But then why use it in first place.

** I should try one of those media sticks that has a USB and micro-USB. But it's not part of this review.

UPDATE: in a small benchmark I copied a file from my laptop to a USB HDD in 1:05 and to a USB flash drive in 7:36. I was able to download the same file from the FileHub USB flash drive in just under 3:30.

Travel Routers Part 4 - Environment

Before continuing with the next test I should mention that I'm doing all of my testing using a Chromebook (ASUS Flip) as part of the minimal hardware needed to get the ecosystem working and since it's my work and home environment.

It was not possible to upgrade the FileHub software from within a Chromebook. The filetype was a rar, which is a favorite of Apple geeks and only recently adopted by Windows, but that was not the raw file format that the FileHub would accept.

Luckily for me ChromeOS understands rar files and so I was able to open and upload the file. The upgrade version number was only a very minor so I'm not expecting anything major... but there you go.

Travel Routers Part 3 - FileHub

Continuing the Chromecast mission I'm trying the FileHub. In hindsight I should have performed a speed test but maybe another day.

  • The initial boot took about 45s; clearly faster than the ZyXEL.
  • It's running under battery power from the factory
  • I did not need to do anything other than connect the main access point which took a few tries because I forgot to enter the password... and so my main complain is that GUI is singularly non-standard, however, adding the local access point did not require a reboot.
  • After the initial boot the WiFi icon blinked blue then solid blue. Once I connected to the access point the icon displayed in green. I suppose this is useful feedback.
The FileHub get's the zero to boot award so far because of it's boot time and config time. It is missing a QoS feature that the ZyXEL has although that's not important in a hotel and neither of them can auto detect firmware updates. (download page)

Same as the ZyXEL I can see the Chromecast working out of the box.

UPDATE: in the hour that I was testing the battery indicator fell from 4 to 3. I have no idea what the scale is so longer tests will be required.

Travel Routers Part 2 - ZyXEL NBG2105

Diving into my ZyXEL travel router I had one mission and that was to get my Chromecast v2 to setup in a hotel. Since I'm no longer in a hotel some features are intellectual only.

  • The initial boot took about 1:12 but shortly thereafter it seemed to reboot. Since it was not connected to the internet it's unclear what just happened.
  • In a hotel I'm likely to be in the WISP+UR mode so I had to change the rocker switch and reboot again.
  • Search as I might the manual was very unclear how I was supposed to connect the router to the existing access point. But then I found the magic: Wireless LAN > Site Survey. From the site survey you select the SSID of the node you want to bridge to and enter the password. The router will immediately reboot and a timer will start.(**)
** The timer is insufficient and the local WiFi will likely redirect to the preferred or one that is locally available so when the travel router completes the reboot you'll have to reconnect to the ZyXEL.

Before I take this router to a hotel I need to change the SSID, default username and password. Since I'm experimenting I'm not sure whether or not I want to name them all the same or if each should be different.

As for the Chromecast, my expectation is that once the travel router is online and ready for normal operation (estimated 7 minutes) the setup will be uneventful.

UPDATE: I should add that the UI is OK but VERY old school.

UPDATE: can not auto detect firmware updates. (latest firmware here) The latest seems to be 06-25-2013 however the info is pretty unclear.

Travel Routers Part 1

The first two of the travel routers I purchased have arrived and while I hate meaningless unboxing videos I'll spare you the details other than to say that the RavPower All-In-One FileHub is minimally packaged both inside and out; just the router and a USB cable leaving me to provide the wallcharger. The ZyXEL Wireless Router included the device, USB cable, universal power br. Both had very little documentation although the RavPower was easier to comprehend and if you want to get beyond QuickStart use-case you will have to download the ZyXEL manual.



RavPower
Battery - 6000mAH (quick press the power button to see power level) While the documentation calls it an "external battery" it seems to also power the router.
Ports - SDCard slot, mini USB for charging the battery, USB port for charging other devices and flash drives, RJ45 ethernet port.
Modes - Access Point, bridge, file server
GUI - web admin, smartphone app
Default SSID - FileHubPlus-XXXX
Default password - 11111111
Default IP Address - 10.10.10.254
Default uid/pwd - admin/(no password)
Manual: link

ZyXEL
Battery - none
Ports - USB power, RJ45 ethernet port
Special - WPS button, Clone MAC button, mode switch
Default SSID - ZyXELXXXX
Default password - 00000000
Default IP Address - (router mode)192.168.1.1 (other modes 192.168.1.2)
Default uid/pwd - admin/1234
Note - There is a Windows driver for autoplay.
Killer feature - From the quick scan of the user manual I discovered that there is a QOS feature. This could be helpful with my Apple Photo Sync problem crashing my modem and routers.

Sandisk Media Drive
- worth a mention but that's it. It was meant as a file server more that a router.

Sandisk Connect
- worth a mention but that's it. It was meant as a file server more that a router.

In part 2 I'll put the ZyXEL and RavPower to the test.

Friday, January 1, 2016

Sandisk Connect

I still own a Sandisk Media drive. I used it a few times when traveling in order to share media with family members as well as proxy the hotel's internet. If memory served there was 64GB internal and an external SDcard slot which I had a 128GB SDcard.



My complaints included WiFi speed, battery life, filesystems were not merged.

This year I purchased a Sandisk Connect 128GB.



My initial impressions were ok, however, in the end I was pretty disappointed.

  • If I set a password it could be reset without erasing the data.
  • Data is in the clear regardless of the password so if I took possession of a drive I can see the data.
  • Connect and data is in the clear
  • No Hidden SSID so when I was in an airplane or hotel I was subject to hackers 
  • Limited power about 2hrs streaming
  • Could not use as travel router as I tried to use it to setup my ChromeCast in my hotel.
Overall not a great experience.

another bad day for open source

One of the hallmarks of a good open source project is just how complicated it is to install, configure and maintain. Happily gitlab and the ...