Rubenerd

This originally appeared on the Annexe, when I documented inconsiderate drivers parked illegally on Mascot footpaths. It used to bug me no end!

Every cloud handles networking between instances differently. Generally you want to avoid using public networks given transit may be more expensive, and has to go through the same public firewall rules your web-facing traffic does.

For Amazon EC2, each instance is given a private IP address which is mapped to an ephemeral public IP, or an elastic IP you reserve. You can tell the private IP based on your hostname:

admin@ip-172-30-0-107:~$ hostname
==> ip-172-30-0-107

Wait, the IP was already there in our prompt. I know the word “literally” gets thrown around a lot thesedays, but that second line was literally superflous. Hey, just like this whole paragraph! Amazon is also a South American rainforest.

With the private IP in hand, you can ssh, rsync, scp etc from it to another over the private network. Right? And yet I couldn’t; I kept getting access denied errors like so many high school dances.

Then I remembered I’d been experimenting with Security Groups to supplement internal firewalls. Unlike these, Security Groups are a blunt instrument applied to all interfaces. By not allowing incoming/outgoing traffic from the private addresses (which would make no sense in a normal external firewall), requests were being blocked - by design.

Allowing SSH traffic from that private interface on the originating server in the destination’s Security Group allowed me to ssh across.

Podcast: Play in new window · Download

40:11 – Toilet paper rolls, Sydney's worst intersection, startup culture, building a data centre, stupid English words, cars in Singapore, ugly new cars with plastic fronts, generic names and slogans, an advertising rant, Mascot construction, listener feedback (Jimbo, Shambles), introversion in children, and the joys of working in coffee shops.

Recorded in Sydney, Australia. Licence for this track: Creative Commons Attribution 3.0. Attribution: Ruben Schade.

Released March 2016 on The Overnightscape Underground, an Internet talk radio channel focusing on a freeform monologue style, with diverse and fascinating hosts.

Subscribe with iTunes, Pocket Casts, Overcast or add this feed to your podcast client.

I saw this ad for a new streaming service:

Michio Kaku’s non—fiction program? How much does this streaming service cost? No wait, I don’t care, take ALL OF IT.

Deepak Chopra’s Factual show? Well, that illusion is shattered.

From Slashdot today:

MIT researchers have created an algorithm that analyzes web pages and creates dependency graphs for all network resources that need to be loaded (CSS, JS, images, etc.). The algorithm, called Polaris, will be presented this week at the USENIX Symposium on Networked Systems Design and Implementation conference, and is said to be able to cut down page load times by 34%, on average. The larger and more resources a web page contains, the better the algorithm’s efficiency gets – which should be useful on today’s JavaScript-heavy sites.

As one would expect from the Slashdot crowd, the highest-voted comment was that we already had this, and it’s called AdBlock Plus. The second one called out NoScript. And they’re right.

An algorithm implemented to reduce page load times would have far-reaching, positive effects. But it attacks the symptoms of bloated pages, not the cause.

I made the mistake of using a browser without NoScript recently, and was stunned at how sluggish every site I went to was. Fonts suddently changing once they’re downloaded. Content injected from trillions of places. Bloated JavaScript libraries and trackers. Auto-playing movies of no value. Maybe most people are desensitised to it, but I found the experience absolutely horrible.

It’s been at least a few months since we had a keyboard discussion. This cannot stand; in part given you’d be sitting to use one.

While I’m okay typing on laptops, I’ve come to use split keyboards on my desktops. There’s some evidence so-called “ergonomic” keyboards do little to protect our wrists from strain, but I do find the angle and position of my arms lead to less pain after heavy days of typing.

This has meant I’ve largely stopped using my beautiful buckling spring Model M clone as my daily workhorse. It also puts certain other classic models out of reach, if we measured typing device ergonomics based on their shelf position.

Take the Sun Type 6, lovingly photographed on the Deskthority wiki. When I first learned Java back in the day, I did so using a Sun SPARC workstation, and one of these gorgeous keyboards. Their rubber domes limited the Ray Charles action one could feel, but the useful shortcut keys and subtle hints of purple certainly left an impression.

Ergonomics aside, their connectors would have hampered their use on a contemporary machine anyway. From Deskthority:

The Sun Type 6 has a fixed cable and the underside has cable channels for both cables going to the left and right. Earlier examples use Sun’s proprietary mini-DIN interface, but Sun later introduced USB on their workstations and servers.

I do remember back in the day searching around for a mini-DIN to USB or PS/2 connector in the hopes of scoring one of the keyboards and using it.

So all one needs to do is make a split, mechanical keyboard in light purple and grey. Anyone game?

If you use GitHub in any capacity, you should be using a public key with your account. It’s one of those blissfully–rare occurrences where security and convenience intersect. GitHub recommends using HTTPS URLs, but I still reckon SSH is easier.

Despite configuring it, pushing commits to a newly-cloned repo resulted in a password prompt. This means the SSH key isn’t being used.

$ git push origin master
Username for 'https://github.com': ^C

This was a new machine reload, and my orchestration scripts had configured Git in the past without any problems. Well, with as few as possible problems as one could expect from Git.

The hint was the “https” in the username field. I’d cloned the repo from HTTPS instead of SSH. Recloning with the SSH url allowed pushing these changes without further prompts.

Lesson learned, don’t tackle Git before your morning coffee.

The Overnightscape Central is a fun weekly podcast hosted by the illustrious PQ Ribber. Hosts and listeners of The Overnightscape Underground participate in a topic each week, and you’re welcome to join.

02:25:41 – Frank Edward Nora!! Shambles Constant!! Rubenerd!! Chad Bowers!! Doc Sleaze!! Clara!! Another multifaceted expanded mix and match set of monologues!! PQ Ribber is your host!!

You can view this episode on the Underground, listen to it here, and subscribe with this feed in your podcast client.

Podcast: Play in new window · Download

58:04 – Quiet neighbourhoods near Mardi Gras, a heart scare, work-life balance, answering Jimbo’s synchronicity questions, placebo beverages, acquired tastes, acquired skills, American comics, tea, nostalgia for American restaurants in Singapore (A&W, Applebees, Coffee Bean and Tea Leaf, Carl’s Jr, Kenny Rogers Roasters) and the world’s most epic muffins!

Recorded in Sydney, Australia. Licence for this track: Creative Commons Attribution 3.0. Attribution: Ruben Schade.

Released March 2016 on The Overnightscape Underground, an Internet talk radio channel focusing on a freeform monologue style, with diverse and fascinating hosts.

Subscribe with iTunes, Pocket Casts, Overcast or add this feed to your podcast client.

There’s absolutely no doubt Australia’s National Broadband Network is a shambles. And not the [good kind of Shambles], which ironically enough would be delivered to your podcast client more affordably and faster with the NBN.

Prior to the last federal election, the Coalition claimed Labor’s Fibre to the Home (FTTH) plans weren’t feasible, and that they could implement Fibre to the Node (FTTN) with less time and money. Rather than having fibre connected to every premises in Australia, fibre would terminate at a node, with existing copper connecting this to homes. Where this wasn’t deemed cost effective, the Multi Technology Mix would deliver services over fixed satellite, hybrid fibre-copper (HFC) and tin cans with string.

It appealed to the broader electorate who bought into Tony Abbott’s Debt and Deficit Disaster nonsense, but those of us in the IT sector knew from the beginning it was technically and financially flawed. and would cost far more than we were being told.

(Unlike some commentators, I don’t fault voters for this. The pressure to discredit the NBN was huge, whether it be silencing ABC journalists or advocating other plans. I place the blame entirely on the Coalition).

It was clear from the start this was partisan politics from the start. Expensive FTTN hardware would have to be built and powered on each street, neither of which would be required in a FTTH setup. Satellite is already in use, and has problems. Since privatisation, Telstra had unsurprisingly failed to maintain their copper networks, which would have to be replaced (and may as well have been with fibre). Ditto the HFC networks that were being bought up to supply NBN. Even new residential developments are getting copper (so-called “greenfield” deployments).

Now leaked reports confirm what we’ve long suspected: the Coalition’s NBN plans not only haven’t saved time and money as a conciliation for lesser service, but are taking more time and money. We’ve strayed beyond absurd into Royal Commission territory here.

This is also on the heels of news that back home in Singapore, customers are getting Gigabit speeds. I’d be more frustrated, but I’d been getting Australia’s NBN speeds in Singapore a decade ago already.