Nginx Requests/Second – RaspberryPi vs Amazon EC2

http://www.evilsoapbox.com/?p=1607

by chris_j • 14 years ago

54 47 14 years ago

CJefferson• 14 years ago

This is interesting. One obvious question, although a little tricky to test. Did the raspberry pi get an advantage by being in the same room as you?

planckscnst• 14 years ago

It would be interesting to try to use netem [1] to match the network characteristics of the link to his EC2 instance and run the test again.

[1] http://www.linuxfoundation.org/collaborate/workgroups/networ...

damian2000• 14 years ago

If I'm not mistaken the performance test software used is running on each machine locally and just hammers the web server with local requests?

jvc26• 14 years ago

Google Cache link - http://webcache.googleusercontent.com/search?q=cache:http://... currently main content not working for me.

FiloSottile• 14 years ago

Already down with 502. <joke> Was the blog hosted on the RPi? </joke>

Amfy• 14 years ago

Yea, I thought the same :)

foxbunny• 14 years ago

So it isn't a joke. Too bad, I thought it was funny.

StavrosK• 14 years ago

Even worse, EC2! :p

stesch• 14 years ago

Worse: nginx/0.8.53

cbg0• 14 years ago

There is a bit of overhead since you are in a virtualized environment on EC2, plus the "micro" instance does not really give you a dedicated processor core - you're sharing CPU time with other instances.

jkat2• 14 years ago

We just finished doing some video encoding testing on a few different platform and EC2 (along with EC2-based offerings) are considerably slower and more expensive. Although 10x more expensive than a 3930K, a cc2.8xlarge instance was only 1.75x faster.

davestheraves• 14 years ago

Where does the 10x more expensive come from? Not questioning just want to understand.

jkat2• 14 years ago

We got a 3930K, 32GB ram with 2xSSDs in software Raid 1 for $169/month.

cjg_• 14 years ago

Comparing cost of a cpu vs price of running an EC2 instance (and over what time)? Bit strange...

dragontamer4• 14 years ago

How about comparing EC2 to a Virtual Private Server? Thats a bit more of an apples-to-apples comparison.

Serverbear notes that Amazon 7.5GB Large instances (which cost $180+ / month) benchmark at ~650 on Unixbench... with 30 MB/s for its disk. In comparison, a 8GB VM from Digital Ocean only costs $80/month. I don't have the numbers for the 8GB VM, but the smaller $20/month 2GB instance has a UnixBench of ~1900 with over 300MB/s I/O from its solid state drive.

(I presume the larger instances have more CPU power / priority in the VM scales)

That is half the cost for triple the CPU performance and 10x better disk performance. Other smaller providers, such as RamNode, offer extremely fast I/O with RAID 10 Solid State Drives in their Virtual Private Servers (500+ MB/s).

Amazon vs Digital Ocean

serverbear.com/239-large-amazon-web-services

serverbear.com/1990-2gb-ssd--2-cpu-digitalocean

To be fair though, Amazon's CPUs are more consistent... consistently bad, but consistent. VPS CPUs and I/O are affected by their neighboring VMs, while Amazon seems to have removed that uncertainty. Nonetheless, in practice, you will always get a better performing CPU and I/O from other providers.

And if we compare both to bare metal servers, obviously bare metal servers win in price/performance, but are harder to maintain, so its hard to do an apples-to-apples comparison. But Digital Ocean VMs can be spun up/down just like Amazon instances... although Amazon has more load balancers and other infrastructure. (But nothing is stopping you from setting HAProxy on a front-end VM to loadbalance a cluster of VMs from Digital Ocean. Even then, other VPS providers like Linode offer Load Balancers as part of their infrastructure now)

Its hard for me to see the case for Amazon's cloud offerings. They don't have very much price/performance at all. At all ends of the spectrum, low end to high end, VPS providers such as Digital Ocean offers more vertical scalability as well as a cheaper price on all of Amazon's offerings.

Unless you need some specialized VM from Amazon (ie: GPU compute), or are locked into their vendor-specific API (oh I feel sorry for you), there is no reason to use Amazon's services IMO.

josteink• 14 years ago

A raspberry pi is by many considering a minimum viable computer of sorts, and the bottom of what one would consider acceptable performance.

Therefore seeing how Amazon compares to that is an interesting exercise. I was personally floored by how poor performance some EC2 instances has for some types of tasks (java/clojure related things among them).

I quickly decided Amazon was not able to serve my needs within the price-range I was willing to pay.

chaz2• 14 years ago

I think EC2 is almost always going to be more expensive than bare metal servers, unless you are taking advantage of the ability to pay by the hour, or leveraging flexible pricing. Reserved instances will save quite a bit of money, but spot pricing can do you even better. Check out this post from a few months ago from someone who is using spot instances for core services, and estimates 70% savings vs on demand pricing.

http://news.ycombinator.com/item?id=4662644

chaz2• 14 years ago

EC2 micro instances throttle down very heavily if you use CPU for more than a few seconds at a time. Other instances allow you to burn 100% CPU all day, but not the micros.

http://gregsramblings.com/2011/02/07/amazon-ec2-micro-instan...

recuter• 14 years ago

Also, with spot pricing you could run a micro instance for 18 months for the price of a Pi and you don't need to wait for it to be shipped to you. :)

RKearney• 14 years ago

You're assuming the spot pricing for Micro instances in the region you pick will not fluctuate which is entirely wrong.

borlak• 14 years ago

You can lose your spot instance at any time if you are outbid...

ck2• 14 years ago

File-System: ext4

Why would you use a journaling file system on a sd card?

Actually I take that back, I was not aware ext4 performance has been improved over ext3 and even ext2

and apparently in ext4 you can even turn journaling off entirely.

raverbashing• 14 years ago

Yes, you can turn the journaling off. You can do that in ext3 as well (ending up with 'ext2')

But maybe JFFS2 or similar 'flash file systems' can be better depending on the use, not sure how the state of these kind of file systems are

simcop2387• 14 years ago

Generally you never want to use JFFS2 or other flash file systems on a flash device that has a separate controller. this is because their own wear leveling will usually confuse the hell out of the controller and can sometimes cause it to either slow down or to lower the life because the controller will not level things properly. Whether ext4 or not is better than say XFS for this, no clue though.

bencevans• 14 years ago

Text only Cached Version: http://webcache.googleusercontent.com/search?q=cache:JCuNKYQ...

simontabor• 14 years ago

Very unfair to be using a t1.micro instance, they're only meant to give very short bursts of power and don't give any dedicated hardware (so giving the hardware specs is pointless).

Do a new benchmark comparison vs a m1.small and it'd be interesting. I bet the small wins by an absolute mile.

CCs• 14 years ago

The blog is called Evil Soapbox. Why do you expect it to be fair? :)

mocko• 14 years ago

I get "502 Bad Gateway". Oh the irony.

dragontamer4• 14 years ago

He is running his site on an Amazon Micro Instance. He probably should switch over to his Raspberry Pi so that he can handle 50% more load.

loeg• 14 years ago

Fwiw, CascadeLink is a high-speed apartment building ISP in the Seattle area (or at least, I that's why I recognize the name). Some of my friends get 30/30 for $40/mo.

mariusbutuc• 14 years ago

Guess what, 502 Bad Gateway!

Nikkau• 14 years ago

A blog is nearly a read-only webapp, it's the easiest type of website to scale.

If you can't avoid 502s with trafic HN can make, you should'nt do benchmarks about hosting matter.

marcuskaz• 14 years ago

the benchmark said the Pi can handle 500 requests/sec, which shows that benchmarks and real world are very different

Groxx3• 14 years ago

While true, it generally means buying bigger hosting / moving to a new provider / some significantly-manual step. If they haven't done it before, it could be hours before it's completed, if they're even aware of it yet.

antihero2• 14 years ago

No, because a static content blog can be served directly from memory. You could probably handle HN traffic on anything.

feralmoan• 14 years ago

Can you turn off rate limiting so we can read the story

Hengjie• 14 years ago

Or just put it on EC2 instances. :P

Nginx Requests/Second – RaspberryPi vs Amazon EC2

Related Stories