Skip to content

Category: Performance

The attack of the killer microseconds

In the Optane article I have been writing about how persistent bit-addressable memory will be changing things, and how network latencies may becoming a problem.

The ACM article Attack of the Killer Microseconds has another, more general take on the problem. It highlights how we are prepared in our machines to deal with very short delays such as nanoseconds, and how we are also prepared to deal with very long delays such as milliseconds. It’s the waits inbetween, the network latencies, sleep state wakeups and SSD access waits, that are too short to do something else and too long to busy wait in a Spinlock.

1 Comment

Gaming Laptops – your recommendations?

The current vacation is hard on me, because I hardly get to use my own computer – the best wife of all and the Schnuppel both compete for time on my machine in order to play Transport Fever and Cities: Skylines. That’s an annoyance not only because I can’t get the keyboard, but also because a MacBook pro apparently sucks as a gaming machine.

So this website lists a bunch of relatively recent laptops with proper graphics cards, and household peace seems to require a premade machine and a transportable device (not a desktop device).

What would be your recommendation (see above, and maybe Elite Dangerous and No Man’s Sky), and why?

16 Comments

BFQ is coming…

LWN reports that the 4.11 merge window opens. Among other things, Maik Zumstrull reminds us, we get

The multiqueue block layer finally has support for I/O scheduling. That is useful in its own right, but the real news is that it enables the merging of the long-awaited BFQ I/O scheduler. That, says block maintainer Jens Axboe, “should be ready for 4.12”.

Of course, if you are on a LTS release of a Linux kernel, it’s unlikely that you will profit from this any time soon.

Leave a Comment

OMG, our cybervaccines are failing

Dark Reading is scared: All new malware is “zero-day”, for an interesting and wrong definition of zero-day, because then the article reads much more impressive.

The actual definition of a Zero Day is a previously unknown exploit that is being used by some party to compromise a machine. In the article, the term is used differently, meaning a file that is a known malware, but has changed itself so that it has a checksum that is not in currently distributed signature catalogs of known malware.

That is of course neither correct, nor new.

Leave a Comment

Load, Load Testing and Benchmarking

(In order to be able to give up the test blog at blogspot.nl, I am moving content over)

So you have a new system and want to know what the load limits are. For that you want to run a benchmark.

Basic Benchmarking

The main plan looks like this:

The basic idea: Find a box, offer load, see what happens, learn.

You grab a box and find a method to generate load. Eventually the box will be fully loaded and you will notice this somehow.

Leave a Comment

Hipsterdoom with Mongobingo

Felix Gessert does a postmortem of the failed Parse startup and product: “The AWS and MongoDB Infrastructure of Parse: Lessons Learned“.

Technical problem II: the real problem and bottleneck was not the API servers but almost always the shared MongoDB database cluster.

And that was with MongoRocks (Mongo on RocksDB) and replacing the initial app in Ruby with a Go implementation of said thing, with WriteConcern = 1, and other horrible presets. All in all, this is like the perfect nightmare of startup architecture decisions.

Felix closes pointing at his current project:

If this idea sounds interesting to you, have a look at Baqend. It is a high-performance BaaS that focuses on web performance through transparent caching and scalability through auto-sharding and polyglot persistence.

Bingo. Also, found the Hipster.

Leave a Comment

The 2017 web is bloated and slow, and I am guilty, too.

Dan Luu has been benchmarking the web, using Webpagetest and other means. Turns out we have given up on optimization – we are doing too many requests, have too many dependencies and our files are just too large.

This site is no exception.

If you are running on Dialup connection or 3G speeds, or add packet loss, everything goes to hell, and most things don’t even load any more.

Only blog.fefe.de soldiers on.

2 Comments

FOSDEM: Graphite @ Scale at Booking.com

Validimir Smirnov gave his talk Graphite @ Scale at FOSDEM.

The slides (PDF) are available for download, and the talk can be downloaded (webm) as well.

Booking stores about 130 TB of data in Graphite, using 32 frontend and 200 SSD storage servers to collect 2.5M unique metric per second,  worth 11 Gbps of traffic in the graphite backend.

This is achieves mostly be replacing all parts of Graphite with API-compatible rewrites in Go and C, all of which are open source.

Leave a Comment