Black Friday '14: E-commerce Pages Far Slower Than They Were in 2013

An anonymous reader writes Black Friday news kicked off this weekend quite early when Best Buy was hit with a massive outage, but it turns out that was only half the story. The top 50 e-commerce websites were slower overall this year compared to last, suggesting customers were frustrated even if they could get to their favorite shopping site. Web performance monitoring company Catchpoint Systems looked at aggregate performance this weekend and compared it to the same timeframe in 2013. The results are notable: desktop web pages were 19.85 percent slower, while mobile web pages were a whopping 57.21 percent slower.

Comments

I did not participate

[SternisheFan]

I don't think a big screen is worth dying for.

Re:I did not participate

[oobayly]

It's only a bargain if you actually need it - not my words, but I did think along the same lines when I was watching the mayhem.

Black Friday scuffles: 'I got a Dyson but I don’t even know if I want it'

Frustrated with not being able to buy a Blaupunkt 40” TV reduced from £299.99 to £149.99, Haggerty rushed to pick up a Dyson Animal Vac, down from £319.99 to £159.99. “I don’t even know how much it costs, I don’t know even know if I’m going to buy it. I just wanted something,” she said. “There are lads in there three, four, five tellies. It’s not fair.”

One of those lads was Andy Blackett, 30, an estate agent, who had two trolleys full of bargains. “I got two coffee makers, two tablets, two TVs and a stereo,” he said. “I couldn’t tell you the prices, but I know they’re bargains.”

Makes me proud of the country I live in.

Re:I did not participate

[istartedi]

Jesus died for your sins. The least you can do is die for a TV. /sarc.

Re:I did not participate

[Skarjak]

What a bunch of fucking morons. Meanwhile here I am reading reviews online to figure out which brand of rechargeable batteries to buy. I want to make sure this ~20$ purchase is the right one. Guess I'm a bad consumer.

P.S.: It's the Eneloops, apparently.

Re:I did not participate

[reboot246]

Are you suggesting that a TV is a sin? If so, you're probably pretty close to being right. :)

Re:I did not participate

[plover]

Clicking "buy now with 1-click" is rarely fatal.

Or did you mean that you didn't participate in the brick and mortar competitive fracas, which has nothing to do with the response times of web pages, which is what TFA is actually about? Even reading enough of the article title to post what you wrote indicates the story is about web pages, which you can't "die for".

Re:I did not participate

[Lumpy]

Glad to see shoppers in the UK are as brain dead as American shoppers.

Re:I did not participate

[LordLucless]

Pounds? I didn't even know Black Friday was a thing in Britain. It's not here in Aus.

Re:I did not participate

[pspahn]

Those Onion articles always crack me up. It's hilarious how close their satire comes to reality ... it's ... erh ... oh.

Re:I did not participate

[oobayly]

Shit, I didn't think of it like that - I've gone from laughing at these idiots to realising that part of my country's* population are a punch line. Next time I hear somebody over here talk about "Stupid Americans" - I'm going to point them to this article.

* A non-card carrying British resident, but my English brother-in-law described me as being more English than he is.

Re:I did not participate

[gbjbaanb]

It wasn't - but it got imported very recently. This year was the first I heard of it over here.

We have the equivalent - Boxing Day (26th December, day after Christmas day) which is when the sales used to start, but for some crazy reason the shops decided to have sales *before* Christmas so everyone popped out and bought their Christmas presents on the cheap. I doubt Boxing day will see the same level of chaos.

Re:I did not participate

[Jane Q. Public]

I bought a "kit" of Eneloops a few years ago with a charger and 10 AAs, and it's still going strong. It has very easily paid for itself, probably several times over.

The Eneloop Pros are even better, they have like a 10%-15% higher capacity. I don't know if that's worth the price.

I want to find some AAAs now.

Re:I did not participate

[wiredlogic]

On the bright side it's nice to have the reassurance that dumb people aren't exclusive to the US.

Re:I did not participate

[AK Marc]

I think I'm pretty reasonable. I was looking for a TV. So I waited until Black Friday. Got a $3000 TV from Sears for $700 (46" 8 years ago or so). Still compares well to new TVs (full HD, but not LED).

Re:I did not participate

[man bear nerd]

Why in the world do retailers let so many bloody people in at a time? isn't there a risk of being sued by those who get injured it is murica? I could never handle shopping like that i am allergic to standing in line to spend money and crowds.

Re:I did not participate

[madbrain]

Just wait until for the deadly delivery drones.

Re:I did not participate

[MrL0G1C]

Get JCBs, long life, high power and don't cost an absurd amount.

Re:I did not participate

[cdrudge]

But we're celebrating his birth by buying TVs we don't need. I'll put off dying until late March/early April (check local lunar calendar for exact date) if you don't mind.

Re:I did not participate

[coofercat]

My favourite related joke: "You know you're working class when your TV is bigger than your bookcase."

In these people's case, it's probably "if your TV has more inches than you have in IQ points" ;-)

Re:I did not participate

[petermgreen]

I read that after panasonic took over they moved the eneloop production to china and quality took a nosedive, don't know if it's recovered.

Re:I did not participate

[bev_tech_rob]

Was needing another monitor for my gaming pc. Checked Best Buy online and they had a very nice 24" Dell Flat Panel monitor for only $99 at Best Buy. Had to work the first half of the day Friday, so went Friday afternoon.

Store was quite busy, but no lines outside or anywhere. Big stack of the monitors in the computer dept. Got what I needed and didn't have to pile onto a rugby scrum to get it. I can't stand that early morning madness. I always wait to later in the afternoon when the idiots have left.

Re:I did not participate

[Cederic]

To me a JCB is this:
http://www.jcb.co.uk/Products/...

I guess you maybe meant this?
http://jcbpowerproducts.com/Pr...

High power indeed, but not sure it's in the desired price range.

Re:I did not participate

[MrL0G1C]

JCB Rechargeable AA and AAA batteries which hold their charge for a year:

http://www.ebay.co.uk/sch/i.ht...

Re:I did not participate

[Immerman]

I did a similar thing for a 40" Samsung for $999 many years ago. Quickly discovered that it wasn't nearly as good of a deal as it first appeared though - I actually ended up with a new stripped-down model made specifically to be a Black-Friday door buster, which lacked many of the features of it's lookalike "normal" model. But while I do occasionally wish it had a full complement of HDMI ports, and have my suspicions that the screen quality isn't fully up to snuff, overall it's been quite satisfactory for my usage as a computer/console gaming screen, especially since I never would have even considered paying twice as much for the normal model. And it's still going strong, so it looks like my old fat-framed clunker of a pre-LED screen is going to last me until quality 4k models get cheap enough to demand an upgrade.

Re:I did not participate

[Immerman]

Hey, still better than those of us who have to face the future knowing that the majority of our country's population are punch lines, and that the politicians embrace that fact with unabashed glee.

Re:I did not participate

[AK Marc]

Strange. Where from? I know Best Buy generally tries to get special deals. I bough a washer-dryer combination there for a lot less than elsewhere, but they weren't directly comparable. The one at Best Buy was XXX-XX-XX-BB and the one at Sears was XXX-XX-XX-SR, or something like that. They code them like a different color, with extra part numbers on the end, do make price-matching impossible. But the units were physically identical. They even had the same part number on the manual and for the manual. But the manufacturer codes were different in stores so that price matching was impossible.

The only problem with my Sharp TV (My only big Black Friday purchase, ever) was that the electrical specs weren't printed by the power port. The manual says is could be "world" or "USA only" power. But there's no way to know. I can plug it into my 240v 50Hz and see if it catches fire. Other than that, they can't say. Seems they used "world" power supplies in some TVs and USA-only in others. Mixed in the same product line with the same part numbers, with no identifiers for anyone to figure it out. From what support told me, even if I opened up the TV, I wouldn't know (or they don't know how I would tell). But that was a standard Sharp feature at the time, not special to the Friday deal.

Re:I did not participate

[Immerman]

Walmart. And yeah, it seems that most everyone runs through stupid no-price-matching shenanigans for large merchants, but there's also the case, like mine, where they release a stripped-down budget model in anticipation of major sales. At first glance it looks very much like the usual shenanigans - same superficial appearance, the usual minor no-comparison variation in model number... but on closer examination is actually a very different model - mine lacked about half the inputs and most of the computationally intensive image-processing options available in the "base model". And like I said, I suspect the panel itself is not quite up to snuff (responsiveness, contrast, etc), though I've since heard rumors that Samsung likes to play bait-and-switch games with the panels in their TVs anyway. I mean bad enough swapping out the power supply with no model number change, but the panel - the most dramatically distinguishing component between TVs... swapping that around just seems like a good way to cannibalize your reputation. You'd think a company with a reputation for quality would try to avoid such things, but I suppose the value of the company's reputation generally isn't factored in to executive bonuses.

Re:I did not participate

[AK Marc]

Yep. Wal-Mart sucks that way. The Levis you get there are badged the same, but are cheaper material than elsewhere. They do the same with lawn mowers and such (there's a book written by a mower maker that turned down Wal-Mart that goes into details).

All sorts of shenanigans are done in the goal of profits. Look at game consoles. They sell for a (small) loss day-1, and replace everything they can that won't break game compatibility with cheaper pieces. Eventually the loss is turned into a profit. Others have learned from that and do the same, replacing everything they can in their toys, as often as they can. That's why consumer desktop models don't restore to the same model. You could buy the first PC of a line and the last, and a backup from one wouldn't restore on the other. But the business lines would try to keep the compatibility up. Turns out that worked the other way. A 5 year old chip costs more than a 6 month old one (because the demand is so low, you have to pay extra so the chip maker doesn't re-purpose the fab). So that's one reason why business computers cost more.

Re:I did not participate

[Luckyo]

High capacity Eneloop-Pros are much worse. The main advantage of Eneloops is the isolation layer inside the battery. This improves both their ability to retain charge over time as well as their expected lifetime to almost a quarter of normal AAs. In the high capacity version, this layer is thinned to make room for more active elements, reducing life expectancy of the battery .

I currently use 3rd gen eneloops (HR-3UTGB). Best AA NI-MH batteries I ever used by far and wide, and I used Ni-Cd and Ni-MH rechargeables since the days of portable cassette-radio players back in 1990s.

Re:I did not participate

[Luckyo]

Eneloops are fine for high discharge tasks. They still retain their main advantage, ability to retain most of the charge over many recharge cycles and years of usage.

Re:I did not participate

[Jane Q. Public]

I did notice on reading further about the Pros that the # of projected recharge cycles was far lower.

By "better" I meant that they have higher capacity. But I did not know of this limitation at the time.

And after having looked at the prices, my conclusion is that the Eneloop Pro is NOT a very good deal. They cost 50%-100% more, for 10%-15% higher capacity and shorter life. I suppose they might be worthwhile if your application absolutely had to have the highest capacity.

In any case, I am happy with my standard Eneloops. Several years now and they are going strong.

Re:But why?

[The New Guy 2.0]

I think it's because FIOS, U-Verse, and Google Fiber all had good years worth of picking up subscribers, so customers want their pages faster and the server-side people didn't upgrade.

Re:But why?

[Anonymous Coward]

Or maybe its because the ecommerce sites didn't pony up protection money to AT&T, Comcast, et al, so they all ended up stuck in the slow lane.

excessive scripts

[whoever57]

Perhaps if those webpages were not laden down with masses of Javascript, doing who knows what, the pages would be faster to load. All that Javascript has to be downloaded from a server somewhere and executed in the browser. It all takes resources.

Many website developers today seem to think that his/her web pages only need to load on the fastest computers as the sole page open in the browser. I think of them as "greedy" websites, because they are greedy with the end-users' compute resources.

Re:excessive scripts

[Anonymous Coward]

And redirects fifteen deep to other sites to serve adds ....

Re:excessive scripts

[tepples]

Many website developers today seem to think that his/her web pages only need to load on the fastest computers as the sole page open in the browser. I think of them as "greedy" websites, because they are greedy with the end-users' compute resources.

Would you say Phil's Hobby Shop is greedy?

Re:excessive scripts

[ihtoit]

yeah I just reimaged an xp box, installed the ITE8212 RAID driver from binary instead of using the reference driver like I usually do - now I got some grey blob in the systray that has no business being there. YES, I KNOW I GOT A RAID INSTALLED, FUCK OFF ALREADY.

Re:excessive scripts

[LordLucless]

The javascript on the primary site I work on takes up about 50% of the page load time. None of it is to do with functionality - it's all analytics or A/B tests or performance measuring stuff. One day something broke with the tool the marketing guys use to insert all that guff, and the site performance doubled. Inspect the DOM tree after it's loaded, and there's 30-50 iframes and script tags that have been dynamically inserted on any given page.

I'm not against javascript; it's useful for making sites do useful things. But this sort of crap just drags everything down.

Re:excessive scripts

[plopez]

And only developed for Chrome "bcaz it's kewl"

Re:excessive scripts

[Razed By TV]

At the risk of sounding like a luddite:
Off and on I will think about how people want to treat the internet as a utility. We try to conserve water, we try to conserve electricity, we are metered for these. For phones, we pay for minutes, though you can opt for unlimited plans, and the infrastructure is such that unlimited plans don't burden others. For cars, we have gasoline (though technically not a utility), and when gas is cheap, people buy larger vehicles. When the price of gas goes up, people become concerned about gas efficiency.

Noone is terribly concerned about "conserving" the internet, or conserving computer resources. Every year computers get faster, and every year websites get less efficient. More bloat, bigger images, more script nonsense. They find more ways to update the browser, to make it smarter and yet more bloated. My NoScript and RequestPolicy plugins are so laden with websites that aren't obviously related to the one I'm on. If I'm lucky, there are one or two sites with a related name, or a CDN, and I can allow these and continue on. If not, I sometimes temporarily allow all, sick of going down the rabbit hole and just wanting to get to my destination.

I'm sure there is an electricity cost related to the extra computing. The time required for page loads is simply time you've wasted, unless you have managed to multitask a few pages. A site taking 15 seconds more than it did a year ago isn't a lot one time, but it adds up page after page, day after day. Even mobile versions of sites, using 4G services, load slower than they did years ago.

I just want my information, I want it simple, and I want it now. I'm sick of all this crap that is designed to make my life better somehow. I liked my life the way it was. I liked being able to do Verbatim searches on Google and actually getting verbatim results. I don't need fancy maps that take 10 times as long to load, I need simple maps that work fast when the network is congested. I don't need functionality changes to make things look slicker. I want to be able to do more with the hardware I have, and we just keep going in the opposite direction.

Re: excessive scripts

[snowsnoot]

I hear ya bro. Just a bunch of asshats with make work projects and ppl trying to get rich. Simple is always better.

Re:excessive scripts

[pooh666]

It isn't the amount, as much as the complexity and the number of different vendors, programmers, companies that are represented. The post a way's up blaming "tracking" is way oversimplified and hitting the wrong issue. How do you debug a page for performance AND MAKE ANY CHANGES, if much of the code you didn't write and you can't change?

Re:excessive scripts

[pooh666]

What is the site? This smells of some poor choices, but mixing in A/B with Analytics in your description is not clear. Are your pages fully cached with a CDN? Are there major content decisions made only in the JS?

Re:excessive scripts

[tlhIngan]

Perhaps if those webpages were not laden down with masses of Javascript, doing who knows what, the pages would be faster to load. All that Javascript has to be downloaded from a server somewhere and executed in the browser. It all takes resources.

Many website developers today seem to think that his/her web pages only need to load on the fastest computers as the sole page open in the browser. I think of them as "greedy" websites, because they are greedy with the end-users' compute resources.

The problem is that developer PCs are often some of the most powerful in the company because they are developers and can demand it. I mean, give a developer, web or otherwise a bog standard PC with less RAM that "average" people have and you'll get nothing but an endless stream of complaints.

So yeah, web developers with Haswell 3.5GHz i7s and 16/32GB of RAM designing webpages. yeah, it loads fast, but bogs down someone with a 4 year old PC and barely 4GB of RAM.

Then there's all the preparation for traffic - yeah, they get all the static CDNs up and running, the database servers are beefed up and the dynamic servers are beefy. Then they forget one script they have on every page references some dinky little server everyone forgot about. That server keels over and the page coding is such that the browser isn't able to render the rest of the page while loading it in the background. So now the pages load slower and slower and slower and everyone thinks it's either the static CDNs, the database or the dynamic views being generated, and not the server handling that one script which is vastly underpowered because it's hidden in the corner of the datacentre and forgotten about.

Re:excessive scripts

[LordLucless]

Most of the JS that causes issues are third party scripts from various vendors, loaded from their sites. If their CDN chokes, it affects our site. All the assets we control are accessed via a CDN, and our pages are cached to the extent permissible by their content. It's the arbitrary crap loaded in from third parties (that can't be cached or handed off to our CDN because it's dynamically generated) that screws stuff up.

Re:excessive scripts

[vtcodger]

> All the user can do is complain that bootup is slower...yeah, y'reckon?

At least one can evict most stuff from the system tray if one works hard enough at it. And it is handy to have a volume control and possibly a few other things there. What, can be done about whackjobs who believe, almost always incorrectly, that javascript is essential to their user's "website experience?"

Re:excessive scripts

[dave420]

So your dev team made some horrific choices, and that's the fault of JavaScript? Interesting.

Re:excessive scripts

[LordLucless]

The dev team didn't make any choices at all. The dev team doesn't write their own requirements.

Re:excessive scripts

[dave420]

They have no say in the matter? Then the company in question is a joke, and to blame anything outside the company when that's the state within is sheer insanity.

Re:The problem is relational databases.

[bill_mcgonigle]

Instead, we need database technology that is new hat. This database technology already exists, and they're called array databases.

Array databases are web scale!.

Too much coding on the pages

[Karmashock]

More of the coding needs to be server side or not exist at all.

The worst is the ads. I turned on NoScript and so many pages just fly now because the stupid javascript isn't allowed to run.

Re:Too much coding on the pages

[sound vision]

I recently installed NoScript for security reasons, but I was glad to find the speed improvements too. Also, the NoScript's domain list has shed some light on how many scripts really are on some of these web pages. They have their own scripts, plus several social networking sites, random CDNs, Google analytics, a couple of ad services... Then you hit "temporarily allow all scripts" and the NoScript list shows even more domains and you realize the scripts are being chain-loaded. Some of these sites end up with 25 domains listed. That means you are waiting on 25 servers to respond, 25 DNS lookups, before the scripts even get to executing, which is even worse.

Re:Too much coding on the pages

[Karmashock]

whatever the reason, javascript and other active code that runs client side is a fucking pain in the ass. I'm making a point of disabling most of it unless absolutely required. It has the added benefit of acting like an adblock in most cases as well.

I use a combination of adblock (only on sites I've never visited before or do not want to patronize), NoScript (because javascript should only be used when it is ACTUALLY needed.), and an Adobe Flash blocker that simply prevents Flash animations from running unless individually enabled.

The AdBlock protects me from getting bombarded with porn ads, NoScript kills the popups and other annoying shit, and the flash blocker stops those fucking flash movies from running on some of the news sites. It is even great for youtube. I like to open five or six youtube windows at once and then play the movies one at a time. If you do it without a flash blocker they all start whenever the windows load rendering the activity counter productive.

Re:Too much coding on the pages

[Karmashock]

Yep. Which is why I keep noscript enabled.

Re:Too much coding on the pages

[Lord Duran]

I find Ghostery to be helpful with that.

.66 seconds?

[Enry]

FTFA:

Median webpage response times for desktop websites for the entire group (aggregate) was 3.991 seconds, compared to 3.330 seconds in 2013.

Do people even notice that? I mean, if I'm getting what I think is a great deal and it takes literally a fraction of a second more for the page to load I don't think I'm going to care.

Re:The problem is relational databases.

[ihtoit]

what would you store this array on?
A key-value filesystem such as ANY DISK BASED FILESYSTEM??

Re:But why?

[thogard]

Page load times are down because pages are loading so many more tracking options and some of them are very abusive on the javascript engines. If you turn on the status line (even if you can as it is gone in some modern browsers), you will often see it saying "loading 159 out of 162" and those last ones never load. There is also something that is related to a compounding latency problem that many developers don't think about it because they don't see it when they are developing the platforms and modern tool kits help to hide it from developers too.

I guess people don't like IBM's old work on the subject that showed dropping a 3 second response to just 2 seconds resulted in substantial improved efficiency. Maybe marketing groups need to understand that a customer stuck on a slow site is a bad consumer.

Blame the ISP

[Lumpy]

All of the ISP's have NOT upgraded their backbone for years, they are now overselling it by never before seen levels making even 500Gb Cable Internet feel like DSL.

The problem is people are not screaming about it to their congress critter forcing ISP's to deliver what they sell. They need to pay a $1000 per user per month fine for not delivering what they promise or advertise.

That would get the lazy executives at Comcast moving.

Re:The problem is relational databases.

[thogard]

If the data is in cache, it is faster to open a file, read the data in as a text file, do a linear search, process the data, close the file than it is to just do a connection to an sql server. Most online stores are selling so few products that their entire product database should fit in L1 cache. If you want fast, make sure everything that has to be run can live in L1/L2 cache except what has to be be written out. Modern file systems are very good at writing out small files quickly without making the person in front of a screen wait.

Re:VMs

[Lumpy]

No we havent. ISP's still heavily oversell.

at 8pm on a friday, I can not get more than 13mbps. I pay for 25mbps. Friends pay for 50mpbs and they cant get more than 13.

ISP's are still dramatically oversold.

Re:But why?

[pspahn]

This is part of it, yes.

The fundamental problem is you have loads of ecommerce sites that were built as turn-key solutions and handed over to an "admin" for the company. They can start creating their own content to add to the site, so they start searching for things to add to their site. They find snake-oil dealers that offer them everything in exchange for a small script element inserted into the DOM.

Additionally, the admins haven't taken the time to learn how to save images for the web properly, and they serve a 900x600 image that's a handful of MBs (x6 for a simple slideshow).

Between the excess of HTTP requests added by the tracking scripts and the excess of MBs being downloaded for images / video, it is not a surprise that ecommerce sites are getting slower. I would expect this trend to continue to some degree.

Re:The problem is relational databases.

[ihtoit]

I can vouch. I had an eshop in 2000 which carried over 3,000 products (RPG merch), I fit the lot, including images, onto a zip disk. The text portion (basically all that was in the database, the images were stored on the filesystem) fit on a floppy. Or, in the case such as mine that I had my own server on a colo backbone, loaded onto a RAM disk. Holy shit, that was one (relatively speaking) quick server.

Re:But why?

[gbjbaanb]

or its because so many e-commerce sites are now hosted on the "cloud" rather than their own servers in a datacentre.

Amazon created their cloud as they had lots of spare capacity in the off-peak so thought it'd be a good idea to sell it to businesses that would use it when the holidays were not on, and Amazon would use it for consumer ecommerce when the holidays were on.

But now, both Amazon has sold its capacity to ecommerce places who need it when Amazon needs it... hence slower sites. I'm sure the same could go for Azure, too little capacity oversold.

Re:VMs

[alen]

except you don't need even that much to surf a web site. virtualization is the problem. all the servers are oversubscribed to use every little bit of RAM and CPU so the bean counters can cream their shorts with higher return on asset ratios

i could surf netflix and youtube just fine over the weekend. amazon was as fast as usual. but toys r us and best buy were like watching trees grow and toys r us store inventory checker was broken

Websites slower in the Cloud?

[lippydude]

I've noticed a number of sites are hosted on some kind of content management platform, presumably aloft in the Cloud. Sometimes there's a noticable delay in getting the page to load.

Re:Websites slower in the Cloud?

[dave420]

So you are presuming they're hosted in "the cloud" and then assume that's what's causing the slowdowns? Holy ass-delving, Batman!

Re:Websites slower in the Cloud?

[lippydude]

Our CDN was unable to reach our servers

Re:The problem is relational databases.

[plopez]

That's just DBA 101, "always buy more RAM".

Re:The problem is relational databases.

[plopez]

I am glad you posted that. I am putting together a little project I call Distributed Integrated Scalable Array Database, DISArray. It will be a shardable web scale instantly consistent DB engine which will have kick ass performance and a Heisenberg query engine support by a look ahead design I have code named "Schroedinger".

Now all I need a is cool mascot and I will be well on my way to becoming a bazillionaire. Zuckerberg better watch out! Look for it on GitHub.

Re:But why?

[lucm]

The real issue is uninformed cloud bashing comments. They are so numerous nowadays that they clog the internet pipes and prevent ecommerce traffic to go thru.

Re:The problem is relational databases.

[Fnord666]

I am glad you posted that. I am putting together a little project I call Distributed Integrated Scalable Array Database, DISArray. It will be a shardable web scale instantly consistent DB engine which will have kick ass performance and a Heisenberg query engine support by a look ahead design I have code named "Schroedinger".

This is going into my next system proposal to management.

Re:VMs

[Lumpy]

The fact that slashdot on it's own requires 10.5mb in freaking Javascript is the other part of the problem.

Low quality web designers that cant figure out how to do something without 600 pounds of "libraries"

Re:But why?

[fahrbot-bot]

I think it's because FIOS, U-Verse, and Google Fiber all had good years worth of picking up subscribers, so customers want their pages faster and the server-side people didn't upgrade.

It's because every web page wants to load a metric fuck-ton of third-party Javascript and Ajax code from 10 different sources - just to display their banner and navigation panels ...

Because you started early

[YrWrstNtmr]

"Black Friday" has been happening and advertised since 1 Nov. Friday Nov 28 just happened to be one day in the stream of pseudo sales between 1 Nov and 15 Jan.

Re:The problem is relational databases.

[tompaulco]

You sound like a guy I work with.He turned isolation level down to dirty reads and believes it will have no effect on relational integrity. It is already causing problems with programs picking up items to work on that are not in the correct status. He believes we can solve all these problems by rearranging code. However, if A and B both must happen for a transaction to occur, I don't see how doing B first instead of A but not doing it as a transaction is somehow going to fix the problem. What if you do B first and A blows up? Now you have data that is out of synch.

Jeez, NewEgg

[rrohbeck]

I gave up browsing at some point, it was so bad. Amazon was a bit slow but worked OK.

Re:But why?

[pooh666]

I don't know how this got moded up, it is nonsense. Most tracking happens post load/post interactive, and someone saying, "looking at your status line" is a telltale for this person not having a clue about website performance opt. I get this from JS devs and PM's all of the time, and I really do have to prove that the 1M of badly optimized images is more the problem.

Re:Custom hosts files complete the equation... apk

[Karmashock]

hey there APK... :-)

Re:But why?

[vtcodger]

It's because every web page wants to load a metric fuck-ton of third-party Javascript and Ajax code from 10 different sources - just to display their banner and navigation panels ...

Sounds right to me. All I know for sure is that today's web has managed the rather remarkable feat of mostly being slower in use than Compuserve was in the early 1990s with a 1200 baud modem. And that's AFTER blocking about 16000 nuisances in /etc/hosts. Our EEE PC's where I don't currently have a hosts file, have become pretty much unusable in Firefox. My esteemed spouse has come to blame Firefox for the situation and would probably advocate execution of the Firefox programmers.

Personally, I think the basic problem is that web site designers are often incompetent and almost universally nuts

Re:But why?

[dave420]

Nope. As has been pointed out elsewhere, the loading of tracking code usually happens after the page has been rendered. This is because the tracking/analytics code doesn't want to affect the loading or delivery of the page, as they are used to measure that (amongst other things).

You appear to just be guessing...

Re:But why?

[dave420]

That would make sense if Amazon only offered their Christmas-peak infrastructure, and no more. As that's not the case, your argument is nonsensical.

It's quite likely the cloud is helping, as now companies can fire up new servers and load balancers to deal with increased traffic in seconds, instead of quickly reaching the limit of their limited in-house, geographically-constrained infrastructure. These servers are located around the world, in places best suited to serve the increased load, giving a real benefit to the hosted sites and the users of said sites.

Or you can just keep on shouting "durrr durrr cloud! durr!" and show everyone you really don't understand how the internet works these days. That'll really help.

Re:But why?

[gbjbaanb]

It's quite likely the cloud is helping, as now companies can fire up new servers and load balancers to deal with increased traffic in seconds

then why didn't they?

The cloud is just as resource-constrained as the old datacentres used to be, only shared across many customers. Its like all those customers joined together and put all their old servers into one big datacentre. "The cloud" doesn't magically increase the number of servers present, it just shares the load. Normally that's fine. But when everyone wants to use it all, all at once, even the cloud goes slow.

Re:The problem is relational databases.

[jbengt]

Isn't the index used to access the array a key value?

Thanks Obama!

[ThatsDrDangerToYou]

(disclaimer: opinions expressed do not necessarily represent my own...)

I'm going to go with Angular/bootstrap FTW

[Pherdnut]

There's a whole new generation of JS devs who are complete slobs about dependencies. They will attach the entire Bootstrap library for one plug-in. I've seen libraries that embedded and minified it such that it wasn't even obvious they were using it and they weren't using it for much. 20 megabytes for a !@#$ing restful documentation widget whose own proprietary code was 20,000 lines long. It's just ridiculous. IMO, every client-side web dev should be forced to support IE6, then mobile, then write for the desktop browser. But people want the latest buzzwords, they want them fast and cheap and they want them now. This is what you get. A bunch of jr-level slobs writing e-commerce apps who knew what to say about how much they love these new frameworks at the interview.

Re:The problem is relational databases.

[userw014]

Perhaps if you have real hardware for your data, and your data is static. But once you escape into the real world, you shouldn't be designing web applications that depend on L1 (or L2) cache.

Cloud

[NewYork]

I'd blame the Cloud

Is the problem the servers or the network?

[TonyXL]

If the network is hitting capacity, why would a network company want to invest in higher speeds, if the gov't is going to tell them how to run their network?

problem isn't with the dev machines

[Chirs]

Those developers may very will need those fancy machines.

But whoever is doing usability testing should be testing on fast machines, slow machines, new machines, old machines, mobile devices, etc. If they're not, then they aren't doing the job properly.

Re:But why?

[Cederic]

Clearly you missed the twenty minute period on Friday in which amazon.co.uk was returning HTTP 502 error codes.

I guess they should've used some cloud servers to cope with the increased load or something, huh?