+972 52-548-6969
Why scalability matters?
time to read 1 min | 42 words
Otherwise, you get this:
And that is one sale that isn’t going to happen!
FUTURE POSTS
- Partial writes, IO_Uring and safety - one day from now
- Configuration values & Escape hatches - 4 days from now
- What happens when a sparse file allocation fails? - 6 days from now
- NTFS has an emergency stash of disk space - 8 days from now
- Challenge: Giving file system developer ulcer - 11 days from now
And 4 more posts are pending...
There are posts all the way to Feb 17, 2025
RECENT SERIES
-
Challenge
(77):
20 Jan 2025 - What does this code do? -
Answer
(13):
22 Jan 2025 - What does this code do? -
Production post-mortem
(2):
17 Jan 2025 - Inspecting ourselves to death -
Performance discovery
(2):
10 Jan 2025 - IOPS vs. IOPS
RECENT COMMENTS
-
By Oren Eini on Answer: What does this code do?
-
By Dmitry on Answer: What does this code do?
-
By Pyth0n on Challenge: What does this code do?
-
By Oren Eini on Performance discovery: IOPS vs. IOPS
-
By kpvleeuwen on Performance discovery: IOPS vs. IOPS
Syndication
| Main feed | |
| Comments feed |
Comments
nah. It's apple, the sheeple will simply accept it and wait. Trust me ;)
@Frans: Not only that, but later the newspaper will report that the demand for their new product was so high that apples servers couldn't handle it, generating more buzz and the impression that the demand must be incredibly high and the product great.
Although I have to admit that the new MacBooks Pros are impressive ;-)
Wish I had the screen shot of HP Site during last year Touchpad fire sale!
What is the point of this post? To point out that a site failed gracefully under a higher amount of load than any site any reader of this blog is unlikely to ever work with?
I can almost guarantee that Apple has scaling systems in place. During a highly hyped conference/product release they experienced a higher traffic volume, likely by an order of magnitude, than they would normally expect, were able to detect it, and display a message to users.
This seems like a win to me. But, sniping can be fun :)
Strange post...
Ironically, your blog loaded EXTREMELY slowly while I was waiting for this post to load :p
In the long term, scalability matters because hardware is so much cheaper than developers: http://www.codinghorror.com/blog/2008/12/hardware-is-cheap-programmers-are-expensive.html
Better throw hardware than throw developer, and by definition, only scalability can allow that!
point taken, but terrible example
I love Apple. I love Apple hardware and operating systems. I believe their web properties are an embarrassment. The apple.com site is a horror to read on the iPhone, heavy and unable to adapt. MoblileMe was bad enough even Jobs made fun of it. But the Apple Store is the worst of them.
Despite being incapable of dealing with order volume every time there's a product release, the site seems to be completely wedded to a synchronous, session-based, 'web object' model that was really ahead of it's time when it was shipped in 1995. It's completely shocking to me that there's a lack of static content fallback, client-side UI, and asynchronous message-based order processing. Don't you lose money when you lose the orders?
....well, I guess not when you already sell as much as you can manufacture. Sigh.
The message is not actually referring to internet traffic - they mean roading infrastructure in the supply chain has been overloaded by people racing to Apple Stores to get their hands on new machines, hence there isn't enough capacity for actual Apple trucks to deliver more product... so they're simply saving you the hassle of 'reserving the unreservable'.
;p
But just curious - forgetting this specific case... is it really reasonable to expect a web-site to be able to service every single human on the planet, simultaneously? Even if they were expecting it? I think it would be naiive to NOT build in a failure case. Whether or not the failure case occurs is out of the hands of developers. The money people will have put the kaibosh on eutopian capacity spending long before....
I think there's just gotta be a point where the business says "If we get that much traffic, then we're probably in such high demand that the customer isn't going to walk away just because they had to wait ten minutes for things to die down..."
"In the long term, scalability matters because hardware is so much cheaper than developers:..."
I disagree so much on that...
David, I run into this issue On January 6th, nothing much was happening then, and it was recurring problem for several days when I tried. It is an _issue_, and while it is fine to say "oh, it happens", it isn't fine when it is repeated offence. And this is a great way to show the correlation between scaling and actual revenue.
CCassio, If you are cheaper than hardware, then that is very surprising.
Maybe yes, I am cheaper than hardware, because I don't use all my time optimizing software. That's a little part of the job of some developers in most of the cases.
But that's not the point. To make software scale you need developers that must know how to make it scale. There are a lot of other arguments, but I'll stop here.
Stop scaling RavenDB and offer cheaper support, so your clients can invest the money on hardware. :)
Comment preview