Random: Birthday, Reflections, Taronga, CeBIT & the Kitchen Sink

#include <stdio.h>
int main (void) {
  printf("Hello World!");
  return(0);
}

It’s been quiet. Too quiet. Sorry for that.

While I did say I was going on holidays and would be on a hiatus, I didn’t expect it to be a complete blackout.

Thanks to my concerned readers checking in on me from time to time. Don’t worry – I’m still alive … but not quite the same person I used to be. Which is a good thing. This post will elaborate a little on the circumstances, before hopping back into the random technical juiciness that regular readers might appreciate more.

Holiday Reflections

After spending nine years at UNSW completing my undergraduate and postgraduate programs, I promised myself a reward of (at least) a year of travelling. With an understanding that this would be the best time to do such a thing as I was not under the pressures of employment, mortgages, or family, I had saved enough from my scholarships and part-time employment with the uni to enjoy myself comfortably.

By late September 2016, the tickets for my first journey were already booked and paid for. The prospect of cheap airfares really solidified the desire to just leave everything and go. At the end of January, I embarked on a ten-week holiday to Hong Kong, Taiwan and South Korea. This was a milestone in my life, as the time spent in South Korea was my first solo trip overseas and a chance to broaden my horizons.

Above: Sai Kung Pier, waiting for a bus to visit East Dam (Hong Kong Global Geopark of China).

While in Hong Kong, I had the opportunity to meet and stay with my relatives, eat some good food, and celebrate my graduation. Chances to go out and explore were not as numerous as I had expected due to the sheer amount of events planned by my extended family. However, I did feel a bit of sadness in seeing the state of the sky in Hong Kong – mostly filled with a mix of fog and smog that wasn’t quite as dense when I visited prior in 2014. Photos of the harbour were pretty pointless given the conditions.

Above: Night market in Taiwan, the first night’s dinner location.

The trip to Taiwan was a one-week country-wide circuit with a Hong Kong-based tour guide with my uncle and aunt on my father’s side. It was a bit of a scouting trip, that resulted in a short stay in each region, punctuated with many moderately long bus rides. It proved to be a good chance to visit some of the attractions in areas which would have been difficult to visit as a solo traveller, and despite being a bit of a “whirlwind”, was still fairly enjoyable. The downside was the sales tactics employed by the tour guide to make up the difference in the advertised tour price and the actual cost of providing the service. While the guide was frank about their business model, some of the red herrings and anecdotes are just a longwinded advertisement in disguise which made things a little less pleasant. Some of the members of the tour group also didn’t help …

Above: APEC House (Nurimaru) near Haeundae, Busan, South Korea – one of my favourite photographs I’ve taken on this trip thus far.

But by far the highlight of the trip was my time alone in South Korea. I really enjoyed travelling on my own, setting my own itinerary, taking my time to appreciate places, using public transport, seeing the locals and at times even attempting to communicate with them. I had many new feelings and experiences – as a person educated in Australia, I’ve never really felt the frustration that comes about from not being able to express one-self because of a language barrier. At one stage, I used my “survival” Korean, a little Chinese (both Cantonese and Manderin) and English to get things done. I also learnt just how funny and problematic translator apps can be, and how useful gestures and actions can be. I was also able to observe history in the making, with the impeachment of former-president Park Geun-Hye, and the protests of the citizens. I appreciated the convenience of decent unlimited LTE connectivity, the various subway networks, a number of driverless subway and monorail systems and their high-speed KTX train to travel between major cities.

When I arrived, Korea felt like a foreign place to me. Despite being an avid K-pop and Korean TV watcher, I wasn’t convinced that it would be easy. After visiting Seoul, Incheon, Busan, Daegu, Daejeon and Gwangju, I’ve learned that Seoul is full of many sorts of attractions and is very well set-up to cater for tourists. As you get further away from the large cities, the difficulty level increases, but even then I was able to manage (to my surprise).

But instead of going into details about the holiday in this post, I’ll save it for when I make the detailed in-depth posts about each section of the trip. This will/might happen when time permits. Instead, I’m still shaking off the “post-holiday hangover” that happens when you return home.

While I was on holiday, I was a completely different person. I ate differently, I lost 6kg, I walked a lot (for me anyway) including uphill treks. I improved my fitness overall, although this has somewhat been lost after being at home. I put in effort to go visit places – almost every day was spent outside from the hotel room, and a lot of time was spent looking at maps and searching for attractions. For once, I slept well and almost on a regular time schedule. I didn’t miss blogging so much, or even catching up on news online, repairing things or even the test equipment at home. The worries of home mostly faded away, with the exception of when the ADSL2+ link at home failed, which cut me off from my machines at home. I did miss the computing power of my desktop machine, as the laptop I had on holidays wasn’t really quite as fast as I’d like, and the screen on it is terrible. But I probably discovered the biggest asset of all – the fact that I am more flexible and adaptable than I thought myself to be. I didn’t really feel homesick in any way, and I gained a new confidence in that respect. I also began to appreciate architecture, art, history, culture and language more as a result. I became more emotionally connected.

When I returned home, I felt the familiarity of home, but also the disappointment of having returned home. It was an odd feeling, riding the train from the airport in Sydney. I felt less safe and less at home compared to when I was in South Korea. I frequently caught myself staying to the wrong side of footpaths, stairs and escalators (keep left here, keep right in most other places). But soon after, the whole routine thing started to kick in, and I felt a desire to do technical things and post blogs again.

The problem was that I couldn’t. Not yet. Because I had a mountain (18,000+) of photos to look through, process, and a whopping 11Tb of recorded data that I bought home with me that deserves attention. It also wouldn’t be me if I didn’t return home to a pile of technical problems that needed attention as well. I also have some further commitments with the uni as well, and other things to look after. These competing things all served to delay my return to blogging, and even as of now, I’m trying to get many different things done at the same time. Some people’s idea of holiday is relaxing … my idea is completely different. My idea of blogging while on holiday? Completely dashed by the fact I was so worn out after each day that I couldn’t think straight. On another note, I’m now starting to think in Korean. Daebak. At least I can read Hangul now.

As a result, while a number of milestones had passed, I wasn’t able to blog about them – the closure of Optus’ GSM network, the move of SBS to H.264 encoding for HD channels, my birthday, etc. In fact, for some of these things, I had collected the necessary data but still haven’t found the time to perform the analysis.

While my viewership has fallen about 25%, oddly enough, whenever I don’t post, the brains at Alexa seem to think my site gets more popular. That just goes to show how “great” big data is (… not very).

Happy Birthday to Me (Really?)

Above: Photo of a Cat from Daegu Arboretum, captioned and posted to Facebook on the day of my birthday.

Another year, another birthday. What’s new? Well, as customary, it’s good to take stock of how the numbers on Facebook are stacking up.

This year shows a decline in the number of messages compared to last year, equalling that of 2015. As usual, most of the messages (except two) are “canned” messages which … may not have required much effort on the senders’ behalf to lodge as engagement.

However, if we look at it as a percentage of the number of friends I have, this year is marginally worse than all years on record. This is due to the increase in number of friends, and decrease in the number of friends that are sending such messages even though the majority are somewhat insincere.

The demise of Facebook? Well, it didn’t quite happen as I expected, because the trends weren’t linear (it was a joke anyway). But the demise of interaction on Facebook is definitely happening.

You can see the desperation that Facebook has in trying to “grab” users to open their app and/or do something. Now they’re even reminding me multiple times that I haven’t updated my profile in four weeks and they’re telling me when someone else posts just because they haven’t posted in a while. They also like to tell me about events some of my friends go to, even though I’m not invited.

I’m almost certain that if I were to update my profile, it would at least generate a box in some of my friends timelines that says “Gough Lui just added x to the profile” or something similar. Maybe they’ll even get a notification about it. In other words, they’re trying to inspire a seed of interaction which will have a multiplicative effect on their engagement numbers. But the truth is probably that most people are tired of it anyway – they know who they want to pay attention to, and that’s about it. The novelty of updates every 5 minutes has worn off. As a result, I even get the “you’ll see more stories on your timeline if you add more friends” message from time to time.

Was that all?

Nope. This year, I celebrated my birthday with Taronga Zoo (again). Last year proved to be a fruitful adventure despite a bit of rain, so this year, I went again. I was feeling fairly tired since I wasn’t able to sleep well that night. The weather wasn’t predicted to be good, but there was a massive cruise liner at the terminal at Circular Quay, which was an impressive sight.

Some of the planned renovations I saw last year have already been completed, with a new theatre opened. However, some of the other areas are still in-progress it seems. This particular lizard was spotted basking in the open, as if it had somehow escaped …

The giraffes are always a good sight. They were enjoying their time, leisurely visiting each of the food baskets in turn.

The chimps were interesting as well, since there’s a young one. It’s a bit of a handful, and it seems that the other chimps are having fun too – biting its finger for some odd reason and keeping a careful watch.

I also got a chance to snap a few more shots – the first time I managed one of the meerkats through glass. The elephant was feeding, although the viewing platform was under renovation this time around. Unfortunately, I ended up leaving the zoo a lot earlier than anticipated as the downpour was quite significant and I was pretty tired. Definitely a good way to spend the morning. Even the fish are showing off …

As a bonus, it looks like the Opal single-fare ticket machines don’t have enough ticket stock within them for high-patronage stations …

CeBIT Australia

Another annual event is CeBIT Australia, a trade exhibition that comes to Sydney around May. I’ve been visiting it on a yearly basis almost religiously, so it was only right for me to visit again as I was in Sydney.

That morning, I decided to train into the city via Yagoona station, which treated me to this unexpected sight of track-workers digging up some ballast on a “live” section of track that’s running commuter services every 15-minutes. Good to see the high-vis and the horn signal of the driver, acknowledging the presence of workers near the track and to slow down.

This year was unlike the last two in the sense that it is held at ICC Sydney. CeBIT had to move to Sydney Olympic Park as the old Sydney Convention and Exhibition Center was demolished for redevelopment. The ICC had been opened for some time, but I haven’t had a chance to visit until now. Coming in via the light-rail stop, I was greeted by a long wall covered in an LED display panel simulating digital bitstreams flying across the wall with a faux kilobyte counter. It’s definitely a little interesting.

However, the way around the ICC seems a little convoluted, involving escalators up and down. The showfloor also seemed a little smaller than in the old Sydney Convention and Exhibition Center days … maybe it’s a bit of a sign of where the Australian market is heading.

There were a number of Vloggers on the show-floor taking their time interviewing people and filming videos, so I’m sure I don’t need to go into much detail, but the feeling this year is that the number of large companies represented at CeBIT AU is reducing. Instead, a larger fraction of the stands seem to be dominated by start-up businesses and young businesses – some with purely “software” platform offerings, others with a little hardware to go along with it. However, most perplexing is the sheer amount of space taken by educational institutions this year – every major Sydney university had representation at CeBIT this year, which is (probably) a first. There was also a number of “innovations” on display which felt a little less ready for market as prototypes or mock-ups without preliminary data. I’m not sure this lines-up with the primary audience of CeBIT, which are probably business IT decisionmakers looking for technologies that can be deployed imminently, rather than technologies which are yet to be proven or yet to exist. Instead, it feels a bit more like a pitch-fest of companies looking for capital.

That being said, I did manage to have a chat with a number of friends in companies that have had long-time representation at CeBIT (e.g. myNetFone, Multitech, Congatec, Interactcard, Icom), some of which have returned after a hiatus (RFI) and newly exhibiting majors (e.g. Maxo, Crestron). It was also interesting to see some Taiwanese product representation for Lian Li, Aver (formerly AverMedia), Edimax, and more. On the whole, however, the number of exciting things seemed relatively numbered.

On the one hand, there was a display from VicHyper about their Hyperloop research which they are quite optimistic about. However, there wasn’t really that much to see except a static prototype vehicle.

Crestron’s networked video solution was particularly interesting for a number of reasons:

For once, it supports 4k at 60Hz. Another is that it doesn’t chroma sub-sample. And it supports HDR, all over a GbE link. Pretty amazing, but also pretty expensive, as I’m assured.

But even more interesting is what powers it – the Intel Arrira 10. This is the first Intel chip that has been produced with FPGA logic and ARM CPU cores in the one package, something predicted by analysts after their acquisition of Altera but not really seen. It gets my interest because of its bleeding edge capabilities, and the solution that powers it.

Another interesting thing was at the Congatec stand, where they are now starting to ship embedded/industrial boards with Optane memory support as well. While I thought Intel had just about killed off the Atom, there are actually newer generations being made, just that consumer devices now rarely adopt it. As a bonus, the chip on this board has Intel Confidential marked on it – that caught my eye.

Aside from that, as expected, battery based energy storage seems to be the next battleground, with a number of LiFePO4 based solutions on display. Despite this, having proper intelligent integration with solar inverters is a necessary step, and one hopefully the industry will be able to solve. The promise of these batteries are quite significant – the traditional flooded/sealed lead acid variety are well “understood” but have lots of cost, maintenance, lifetime and logistics issues. The newer batteries are much lighter for the same stored energy and do not suffer such depth-of-discharge penalties, making it much more attractive. Maintenance requirements are also generally less, but lifetime and safety are only assured by having a sophisticated battery management system.

So yes, another year, another CeBIT. I can’t help but leave feeling a little deflated, as it seems the show has gotten less focused and less exciting over the years. Maybe next year will be better.

From the Sink Accident Investigation Bureau (SAIB)*

* not a real investigation bureau

Two days prior to the accident, a puddle of liquid of unknown origin was observed pooling near the rear doormat of the garage area. The liquid was clear, odourless and suspected to have leaked from a nearby container, although conclusive evidence was not available. All nearby containers were drained of liquid and/or relocated and the situation was monitored.

Further monitoring a day prior to the accident revealed that the liquid did not completely evaporate, and the ground remained damp although less wet than before.

On the morning of the accident, the mat in front of the kitchen sink was found wet. Upon later hand-washing dishes, water was found issuing from gaps in the cabinetry underneath the sink. Opening the door revealed the lower cupboard completely waterlogged, with water leaking from behind a duct-taped inspection eyelet.

After removing all items in the cupboard and removing any visible traces of liquid, the U-bend trap was disassembled and cleaned. Copious amounts of hair-like material with a green-black mould were observed, along with a pair of complete chopsticks. It was determined that the eyelet cover was missing and improperly substituted with duct-tape contrary to manufacturers’ instructions.

The trap was reinstalled, with a whittled cork bottle stopper as a replacement eyelet cover, wrapped with a vinyl glove finger and rubber band as a secondary line of defense.

The most probable sequence of events are as follows:

  • Prior occupants may have had an under-sink water dripping incident, which was caused due to obstructions accumulating within the U-bend trap causing the inlet water level to rise above that of the inspection eyelet especially under heavy flow.
  • The whereabouts of the eyelet cover were not able to be determined, but it is presumed that it may have been blown out due to static pressure of the water column and lost.
  • Instead of clearing the cause of the leakage, the prior occupants performed an unauthorized repair involving the use of duct tape to stem the flow of water.
  • While initially effective at forming a seal, the adhesive failed through over time. A slow leak formed, however, due to the accumulation of matter within the eyelet, the flow rate was limited.
  • Liquid flowed through the gaps in the lower shelf made for plumbing and wastewater thus entered the slab and migrated to the wall facing the garage, where water slowly permeated through to the floor at the garage due to its lower level.
  • Due to the slow rate of the leak, this was not noticed until the leak grew in intensity due to the increasing level of blockage within the U-bend trap.

To improve the safety of sinks, it is recommended to:

  • Inspect under-sink areas to confirm the absence of water, and any inspection eyelet covers are correctly fitted and sealed.
  • Periodically clean any U-bend traps which exhibit symptoms of clogging.

Report signed Gough Lui for the SAIB*.

To Upgrade or Not to Upgrade

We seem to be in a particularly interesting time for computing, as it seems to have been one of the longest periods of time I have gone without an upgrade to my day-to-day machine. It seems the sufficient capabilities of many older machines, coupled with more reliable components and better utilization methods have resulted in the ability for many consumers to defer their upgrades significantly.

In the past, upgrading every 2-4 years was considered normal. This year, the core of my AMD Phenom II X6 1090T which I got second-hand from a friend who wanted something better has turned seven years old. In the five or so years I’ve been running it, it has been pushed to 3.9Ghz (formerly 4Ghz, but it failed to maintain stability after 2 years of operation) and never skipped a beat.

Even the monitors I’m using are 16:10 aspect ratio (a since extinct aspect ratio) of a TN LCD panel with CCFL backlighting, all no-yonger than about 7-10 years since manufacture. All of them have rolled over their service menu time counters in some way, but my BenQ G2000W has just registered over 3 years of screen-on time. That’s pretty amazing.

Not to be outdone, I have my Samsung F4 2Tb hard drive that was used as a file-store drive that survived several servers. It has just ticked over 51k hours (5.9 years of continuous operation), with still mostly-healthy SMART vital signs and no signs of any data loss.

I’ve always been rather careful with my money, so I have a tendency not to replace equipment until it truly fails – and by that I mean beyond reasonable repair or work-around. My mouse, for example, has a broken scroll-click button which has gone intermittent. As a result, I’ve remapped it to the thumb back-forward buttons, and developed a new habit rather than replacing the mouse.

The problem is – at some stage, it makes sense to replace things purely from a technological performance perspective. For example, 4k IPS monitors are common enough – even though they’re not cheap, I’m sure I’d enjoy editing my photos a lot more with them. The new AMD Ryzen (and soon, Threadripper) CPUs are going to give me about 4x the performance (or more) of what I have now – so chances are I won’t have to wait for Lightroom or Handbrake as often as I do now. Support for 64Gb of RAM means less swapping-out to SSD, as I only have 16Gb now as the maximum supported by the motherboard. PCIe M.2 NVMe SSDs offer data rates about 5x as fast as any SATA3 based SSD, with comparable capacities as well.

The temptation is real. After all, Ryzen is pretty amazing. The expense is real too, sadly. If I were to satiate every desire, I could probably afford a small car. After all, my present machine is not particularly “light” on hardware – having every SATA port occupied by a drive of some description is normal, and getting some new ones purely for reliability reasons wouldn’t be unusual.

So what am I to do? Well, for now, the answer is just not to upgrade and live with it. It’s not the best machine, but it’s still enough. The biggest issue with upgrading is being forced to Windows 10. Don’t run Windows 10 and you’ll need to resort to a sketchy hack to maintain your Windows Updates, and still have to deal with a loss of USB performance thanks to no UASP support, and a loss in graphics performance for gaming due to the lack of support for the latest DirectX. A loss of legacy interfaces and software support is also annoying.

But I figured I could make my life a little better – the Gigabyte 890FXA-UD7 that I’m running on has a 16Gb RAM limitation according to Gigabyte. As a result, many years back, I bought 4x4Gb DDR3 1600Mhz sticks to fulfill its potential. However, I noticed some people running 32Gb RAM somehow. As a result, I pilfered two 8Gb DDR3 sticks from two systems and swapped out two of the 4Gb sticks instead to give 24Gb total.

The results were interesting. With two 8Gb sticks in slots 1 and 2, and two 4Gb sticks in slots 3 and 4, the board reported 32Gb RAM, 24Gb available (under Windows). With the 8Gb sticks in slots 1 and 3, and the 4Gb sticks in slots 2 and 4, we got 28Gb RAM with 24Gb available.

The trick was to put the two 4Gb sticks in slots 1 and 2, and the two 8Gb sticks in slots 3 and 4. Then it came up as 24Gb total/usable.

This suggests there is a bug in the BIOS where depending on how the RAM is installed, it is presented not contiguous in address space, with the “holes” marked as hardware reserved instead (or similar).

Unfortunately, as before, the Phenom II x6 1090T cannot run all four sticks at 1600Mhz, so I had to settle for 1333Mhz thanks to the weak memory controller. I did tweak the timings a bit, but then got into trouble when I tightened them a little too much. Thus I needed to do a CMOS reset, but then the board was stuck with the debug LED at 25 and refused to boot.

After fiddling about, it almost seemed that my PCI sound card (a Soundblaster XFi Xtrememusic) had “fried”. Remove the card, system boots. Put the card in, system fails to boot. Except for that was not the right conclusion.

Instead, I discovered another quirk of the F.5f BIOS – on reset of CMOS, the default graphics initialization is PCI slot first. It should have looked at the PCI slot, realized it was not a graphic card (as it would have had its function type as multimedia, subtype as sound) and moved onto the PCIe slot. If you boot first, set the graphics to initialize PEG first, then shove the card back in, it’s all fine again. Cost me a bit of time to work that one out.

But at least now, I have an unsupported memory configuration that works, has passed stability testing and makes life a little more tolerable. Will I make it 32Gb? Probably not. Not because I don’t need it – but because it seems silly to invest in DDR3 when DDR4 is already mid-life. Just not good value for money.

Conclusion

In true random-post fashion, it’s been a bit of a catch-up of sorts. While I did want to post while away, that never happened. Finding the time to do so when home is not easy either. The problems still keep coming in, and the solutions also keep rolling out – the time to document them is what is missing. I hope there’s something for everyone to enjoy – I’m definitely still alive and kicking.

Hopefully, in the future, once I have my act together, detailed postings with holiday photos will start to surface. But before then, there are a few different technical posts to come as well, so stay tuned. There’s also the Vivid Sydney light festival, which I should probably visit and take some photos of.

However, also be prepared for another period of blackout, as I have an 11-week holiday that has been booked and paid for, starting in just under two-months time which will see me visiting Singapore, Japan (Tokyo, Osaka, Kyoto) and Hong Kong. I hope it will be just as exciting and interesting as my first journey.

About lui_gough

I'm a bit of a nut for electronics, computing, photography, radio, satellite and other technical hobbies. Click for more about me!
This entry was posted in DIY, Event, Opinion, Photography, Travel and tagged , , , , , , , , . Bookmark the permalink.

6 Responses to Random: Birthday, Reflections, Taronga, CeBIT & the Kitchen Sink

  1. Thanks for the pics & the SAIB report.

    I’d have to disagree with you on the DDR3 memory – memory is still pretty cheap (in the US, anyway), especially used memory. However, if you’re on holiday, you won’t be using it anyway.

    • lui_gough says:

      I suppose cheap is relative. 8Gb DDR3 sticks are rare to find, especially used, since most people never really opted to max-out their machines. For a nice set of 4x8Gb sticks, it’s probably going to set me back close to AU$300 (or US$225-or-thereabouts). When you factor that’s about half the cost of a new Ryzen 7 CPU or indeed the whole cost for my present smartphone, it seems “expensive” in that sense. Considering I probably won’t get that much use out of it (seeing as the board itself is 7-years old and could fail at any time), I’d rather put the money towards some more modern machine in the future (or even DDR4 RAM if I had any use for it, since it has a chance that a future upgrade could re-use it).

      It has been rather surprising just how reliable my present-day machine is … the Phenom II x6 1090T at 3.9Ghz really pulls about 160W through the VRMs – moving to 4.1Ghz causes the VRM overcurrent to trip. I didn’t expect it to last even four years at the “edge” but here we are.

      – Gough

  2. sparcie says:

    Nice to see you back. I look forward to seeing the holiday photos.

    I’m also running old hardware on my main desktop, although it’s a bit older (about 10 years old now). I’m running Debian Linux at the moment to extend its life a little more (using wine for windows apps). I’ve held out on an upgrade for ages for similar reasons to yourself, basically my machine has been adequate for my needs for quite some time. I’ve been reluctant to decide on a new machine, but the new Ryzen systems are quite tempting. It’s good to see AMD starting to make more competitive chips again. I’m not sure whether I’ll go with windows 10, I use it on a work machine and it seems fine, but backwards compatibility could be a problem if I want to play my older games (which I’m still very fond of) the alternatives are linux which has compatibility issues of its own and simply maintaining an older machine like I do for DOS games. I don’t have any win7 or 8 licenses so I don’t have them as an option.

    I don’t think the Crestron gear uses GBe strictly speaking (although it does use the same cables with the same termination) It’s a standard known as DM which stands for Digital Media. It is used to interconnect their devices including things like video matrices. It is better than using HDMI/DP as you can have longer cable runs, you can send stuff like serial through it as well, and it integrates well with their other equipment as it is used as a standard for all of them. I have a bunch of their gear installed at work, and it generally works very well.

    I didn’t mind the black out, you should after all be enjoying your holiday.

    Sparcie

    • lui_gough says:

      Nice to see you again too :).

      Indeed, Windows 10 might seem okay at first, but the more the updates change things, the more I get frustrated with it. At times, after major updates, application preferences are reset to defaults. I’ve had a few apps really play up and refuse to properly run under Windows 10, but I think the privacy implications and enforced updating are a big reason for my dislike. I am very much an AMD supporter – I hope this brings a new era of innovation and competitiveness, as Intel certainly haven’t been bringing much to the table on a generation-by-generation basis.

      Older Crestron gear were not all real IP over Ethernet – but this one apparently is as it does support 802.11x authentication/encryption, and I was assured it uses IP Multicast, rather than just using the relevant cables with their own signalling or Ethernet but not IP. The way it works (apparently) is that they use JPEG2000 frame-by-frame encoding, so it’s not as bandwidth-efficient but much better on latency and supports 4:4:4 60fps HDR as a result. It’s a bit of a computing power to storage trade-off that is so common, but apparently, having the FPGA section of the Arrira 10 was the “saving grace” that made it all possible. Of course, expensive premium gear not meant for the average home consumer, but always nice to drool over the “real” industrial stuff.

      I’m hoping to have a few moments to shove up a few more posts of short reviews, various findings, etc. Sadly, right now, I’m battling the flu (again)! Aside from that, I’m also battling occasional (10-20 minute) outages a day thanks to all the crawlers hammering the site at the same time (e.g. BingBot, GoogleBot, Yandex, etc) resulting in something crazy (50+ requests per second). But short of emptying my pockets for what is admittedly, a hobby site, there isn’t much I can do. I’ve already spent enough time .htaccess blocking some smaller ill-behaved bots, and leveraged several tiers of caching … as long as it doesn’t go down while I’m writing a post, I’ll be satisfied :).

      – Gough

      • sparcie says:

        That’s interesting with the Crestron stuff. The gear I have is probably about a year old now, so it’s fairly young, but it still uses DM. Not that switching to gigabit ethernet is bad, it’s a great idea as it will be more inter operable and can co-exist with existing network infrastructure. Providing they can make it secure enough. Sounds like it needs a bunch more compute power though.

        It seems like it is that season, I’ve got a bit of the flu as well. Not much fun.

        Probably not much more you can do about the crawlers, I thought cloud flare might have taken more of the load off.

        Sparcie

        • lui_gough says:

          I suspect they will still maintain their “other” system in parallel, as there are places where the lower cost is attractive and whether it uses Ethernet+IP is not a big issue.

          As for CloudFlare – it’s doing a pretty decent job in keeping things afloat, but it’s not particularly easy when you’re doing 400+Gb a month of served content from bottom tier shared hosting. I can’t say I expected more than this, but at least now I have a decent 500 ErrorDocument so that users understand the situation better along with CloudFlare’s “Always Online”, so the page will be served from their cache (if available). Unfortunately, despite careful setting of expiry times/Etags on resources, CloudFlare doesn’t seem to cache as well as I would have liked lately – very likely as paid customers are probably prioritized and their caches may be under pressure resulting in my content being evicted sooner than in the past. Still miles better than nothing though.

          – Gough

Leave a Reply to lui_goughCancel reply