Advances In Hirsute

Since I first launched Hirsute, I've been plunking away at it, making little changes here and there. I thought I'd do a quick post about the changes, some of which I'm quite happy with.

Specifying Histograms

One of the main things I wanted out of Hirsute was the ability to generate data based on non-uniform histograms. For instance, if most of your users have 0-10 friends, some other percentage has 10-50, and a small amount has 50-100.

But specifying that distribution was non-intuitive. You had to create an array of probabilities, they had to add up to 1, and they had to be the same length as your buckets.

Pondering how I might make it easier, I realized that what I wanted to do was draw out the histogram and let the system figure it out. So that's what I did.

This is now valid:

star_rankings = <<-HIST

****

********

**

HIST

and then you can add a generator as follows:

one_of([1,2,3],star_rankings)

Histograms no longer have to add up to 1 — the system will scale values appropriately — and they can be different lengths, though a histogram with more probabilities than values will throw an exception, while a histogram that has fewer probabilities will generate a warning.

Ranges As Results

If your generator returns a Ruby Range object, Hirsute will return a random value (based on a uniform distribution) from within that range. That lets you easily construct a script for the friends example above:

one_of([1..10,11..50,51..100],[0.75,0.2,0.05])

MySQL Batching And CSVs

The MySQL outputter now bundles up inserts for faster loading. CSV is now a supported output format.

Post-Generator Blocks Run Within Object

When you attach a block to a generator, the code in that block will run within the context of the generated object. This lets you access existing fields within the newly-minted object.

All Together Now: A Look at Concurrent Languages

Over the past several months, I've been taking a look at various languages that advertise easy concurrency and scalability. It's too late to use any of them on SimCity, but I'm always thinking about what I'll build next and how I'll build it, and these languages are on my radar. Java is increasingly cumbersome to me as a language, and its concurrency constructs are too error-prone even for senior programmers, not to mention that the thread model Java exposes has serious performance issues if not carefully managed.

Here's my quick take on the three languages I focused on: Erlang, Scala, and Go. The TL;DR version: I'd use Erlang for infrastructure in a heartbeat, Scala will make your developers more productive at the possible expense of application performance, and Go is fast but less fun to work in.

For each language, I worked through at least one book on the subject and then built something for myself. The personal projects ranged from small to sizable.

Erlang

Erlang was my favorite of the three languages, and it would be hard to ever argue against using Erlang for back-end infrastructure pieces that require high scalability. Say, message queues, or chat systems, or NoSQL databases, or the backbones of prominent first-person shooters.

Like all these languages, it has a high-level abstraction for concurrency, but, unlike the others, it easily supports passing messages between machines, which bodes well for a cluster of servers. It has extensive fault tolerance mechanisms, even across machines, allowing for robust systems. It has support for hot code swapping, opening the possibility of upgrading a system while it's still live and reducing maintenance windows to nil. It has great support for extracting values out of binary data, which is invaluable when dealing with network traffic and proprietary data formats. It's a mature, proven technology. And it has the benefits of functional programming: more concise code that reduces the number of potential bugs and immutable objects that prevent weird thread-safety issues.

On the other hand, I can probably count on one hand the number of other Erlang programmers I've met. And it's not like taking a C programmer and teaching them Java; functional programming is a distinct mental shift from imperative programming, and it can be hard to get your head around it. That means that you can write your Erlang code all you want, but what about the people who will have to maintain your system beyond you? Its small community also means that while there are certainly lots of third-party libraries for it, it's not the vast universe that Java enjoys. And while immutable objects are easy to work with, they're also expensive because new ones are constantly being made.

Scala

While Twitter's Scala School argues for treating Scala as a separate language, it's hard not to compare it to Java, since it compiles down to Java bytecode and runs on the same virtual machine. And in terms of developer productivity, Scala rockets past Java in my book.

As an application layer language, it has tremendous advantages. You can accomplish complicated tasks with much fewer keystrokes. You can use functional paradigms and immutable objects, but also use imperative style and mutable objects if you need performance, or, crucially, if you need to bring another developer on board with your system. You can enjoy the same concurrency abstraction Erlang provides. You can leverage Java's seemingly infinite supply of open-source libraries. You can incorporate it into an existing Java application, giving you the ability to bring it in without rewriting everything. You can even easily build internal DSLs with it to make your system more expressive and easier to maintain.

But, in my own experiments and in anecdotal evidence, it suffers from sluggish performance. All that great functionality makes your developers more productive, but potentially at the expense of speed. This makes sense; all that pretty code needs to be contorted and converted into Java with who-knows-how-many object creations along the way. Obviously Twitter and Foursquare manage to be fairly fast, but how much engineering time is spent to get them there? On the other hand, a system that enjoys greater and easier concurrency than Java might be more scalable and have more consistent performance under load, even if any given call could be faster in another language.

Go

Go is Google's attempt to build a better C, with a focus on developing distributed systems at a scale that Google needs. Its concurrency model is distinct from Erlang and Scala's, preferring the notion of Communicating Sequential Processes to the Actor model, but neither is particularly superior; each has strengths and weaknesses that fit different situations.

The big win with Go is its speed: Go programs are compiled down to machine code. And while the built-in garbage collection probably means that C would win a horse race between the two, Go is a much less error-prone language to work in. Its community is still young, but it seems eager to improve the language, and a wide variety of useful go libraries already exist. It's hard to compete with the many years of robust Java libraries out there, but Go nuts seem to have filled in the most obvious needs.

I have to admit that I dislike working in the language itself; it lacks the cleanliness of Erlang and the depth of Scala. But there's no denying that its concurrency model is easy to work with, and the programs that you create are nice and zippy relative to their Java counterparts.

Others

I've yet to dig deep into Clojure, though it's the obvious next one. I figure if I'm going to be a fan of functional programming, I might as well go into crazy Lisp land. But I'd worry that it would suffer from the same performance problems — for the same reasons — as Scala.

It seems funny to mention node.js in this post, since in some ways it's all about zero concurrency: a single thread of execution is all you get. Of course, under the hood there's lots of asynchronous work, but it's tied directly to the operating system's I/O. We use it for a subsystem in SimCity, and it, like everything, has strengths and weaknesses. It can do lots of I/O tasks concurrently. Lots. But it's very sensitive to slow code, since that code will block the entire thread when it runs. It appeals to the game developers on my team, since single threads, event loops, and performance-sensitive code are the norm for them.

However, it's not very mature, and the libraries for it can be buggy and incomplete at this stage. I think we made the right decision switching our SimCity subsystem to node.js, since it outperforms its Java predecessor by a long shot, but it hasn't been simple or without issues.

@AsyncToExecutorService

When I first discovered Spring's @Async annotation, I thought it was a great idea. Slap an annotation on a method, and that method would be turned into a task on a work queue serviced by a different thread. A large number of tasks in a web server can be done asynchronously with respect to the incoming request, which means you can respond more quickly to your user. (These days, I'd write a system around events and ignore threads, but that's not the environment I'm in.)

But after a couple of months of using @Async, I found it annoying. It's applied with Spring AOP, which means it can only be applied to public methods on top-level beans and won't take effect on intra-object methods. It also seemed to inevitably cause circular dependency issues. And, perhaps most annoying, all @Async methods go into a single work queue. Together. With no priority, no distinct properties for different kinds of jobs, no control.

While this last issue has been dealt with in Spring 3.2, that's not what I'm using (though support for Servlet 3.0's asynchronous requests is a tempting carrot). And it still has the Spring AOP limitations.

I moved us off of Spring AOP a few months back in favor of straight AspectJ, which has been invaluable. We can declare control-flow pointcuts, pointcuts on private methods, and more. And during that migration I read AspectJ in Action, which features an example that basically does the same thing as the @Async annotation.

So what if I just rolled my own asynchronous execution aspect that allowed me to specify a thread pool to receive the work?

On a quiet morning, I did just that. I started with the example in the book, and then added a couple of my own twists. My @AsyncToExecutorService takes the name of a Spring bean that is an ExecutorService, and routes the join point to it within a Runnable. If you specify an invalid bean name, it throws a runtime exception. I also added some flags so you could declare that you need to run under the aegis of Spring's @Transactional annotation and whether you could be pointed to a read slave.

While Spring's XML files can get a bit wordy, it's also handy to be able to construct a large range of objects directly from that XML. When I want to define a new thread pool, I do it completely in the XML. I also set up our metrics system to automatically grab any ExecutorService beans on startup and record metrics about their current queue size and active threads so that the thread pools could be easily monitored.

I did the work because I needed it for a performance improvement, but I've yet to check in that optimization and have already heavily leveraged my new annotation into several other areas. My current favorite is a thread pool designed to discard tasks when its backing queue fills up. Non-critical tasks go into this queue and, if we're under load, they just start getting tossed. It's a built-in safety valve.

Hirsute: Fake Data for the Real World

At a previous job, which sold web services to help school districts track their tests and their progress against state tests, the sales team thought they could be more effective if our demo district was more realistic. The demo district, which had survived years of service, only had a few hundred students. Superintendents wanted to get a sense of what the system could really do for their needs.

You couldn't just show a district some other district's data, though, for confidentiality reasons. So one of our engineers came up with a solution. He took data from other districts and munged it together. He didn't just swap names around, though. Because we tracked demographic information, he squished students together within demographics. So you'd end up with Hispanic names together. And those students would have test scores that mirrored the Hispanic population in your district (since, at that point, we probably had data from nearby districts in your state). Kids of different socioeconomic status would also have similar test scores, and so forth. I seem to remember that one district half-jokingly suggested they just run the district off the demo they saw, since it was so close to their own.

I thought of that recently when working with our loadtesting group. I'm sure EA's centralized loadtesting group is no different than that of any other corporation in that they lack intimate details about our business objects. On Spore, I remember that our loadtesting database was set up with something like 100,000 users (we have 3 million or so now), each of whom had something like 10,000 creations (most users probably have 20). Or all the sporecasts had 5,000 items (most held somewhere in the few tens of creations). Early work on SimCity's system produced similarly unrealistic data. That kind of data makes it hard to tune queries and indexes, figure out caching strategies, and all the other normal stuff one needs to do with data for a system.

What I wanted, I thought, was a DSL that would let me do what we did for school districts: Specify how the data should kind of look, and let the system generate that data for me.

Thus, Hirsute was born. Hirsute is an internal DSL built on top of Ruby and its extensive metaprogramming facilities (an aside: I recommend Metaprogramming Ruby for its solid information, though most emphatically not for its trumped-up dialog and narrative structure). In Hirsute, you build templates that define how objects should look, and then you create collections of objects derived from those templates. You can specify histograms of probabilities so that you don't just get a random distribution among options but a distribution that reflects your real-world requirements. You can read in data from files and randomly combine them to get plausible composite strings. You can then flush the collections out to files ready-made for loading into a database (mysql at the moment).

For instance, here's some Hirsute code from a fictional wine-cellar-management service that I created as a sample (since the data requirements most on my mind are for SimCity, which I can't talk about). There's also a manual that goes into greater detail.

# This defines a tasting note that a single user might write about a single bottle. It pulls descriptors from various files.
a('tastingNote') {
  has :tasting_note_id => counter(1),
      :description => combination(
         subset(
           read_from_file('wine_cellar_aromas.txt') {|text| text + ","},
           read_from_file('wine_cellar_aromas.txt') {|text| text + ","},
           read_from_file('wine_cellar_aromas.txt') {|text| text + ","},
           read_from_file('wine_cellar_aromas.txt') {|text| text + ","},
           read_from_file('wine_cellar_aromas.txt') {|text| text + ","},
           read_from_file('wine_cellar_aromas.txt') {|text| text + ","}),
         subset(
           read_from_file('wine_cellar_flavors.txt') {|text| text + ","},
           read_from_file('wine_cellar_flavors.txt') {|text| text + ","},
           read_from_file('wine_cellar_flavors.txt') {|text| text + ","})
         ),
      :rating => one_of([1,2,3,4,5],[0.1,0.1,0.4,0.3,0.1]),
      :bottle_id => 1, # filled in later
      :user_id => 1    # filled in later
    is_stored_in 'tasting_note'
      
}

tastingNotes = collection_of tastingNote

That sample defines a template for tasting notes. The description field comprises 1 to 6 lines pulled randomly from a file of aromas combined with 1-3 lines randomly read from a file containing wine flavors, all joined to one another with commas. The rating is from 1 to 5, but weighted such that most wines will have either a 3 or a 4. The tasting_note_id is a counter that's incremented with each new object. The bottle_id and user_id fields are filled in later when an actual tasting note is created.

Then you define a collection to hold them. You could also do this by using

   tastingNotes = tastingNote * 6000

Which would create 6,000 tasting note objects based on the formula you provide.

So far, the system is pretty simple, but it gets the job done. And because it's a Ruby DSL, you can always just write raw Ruby to fill in what you need. I definitely plan to keep adding to it, though, with a newborn in the house, maybe not quite yet.

Scripting Campfire, Again

A little over two years ago, I wrote a post about scripting Campfire, the group chat tool from 37signals. At the time, my script posted a routine "today's date is" message with a variety of statistics. Over time, the statistics have disappeared — though I now post charts from Graphite — but the bot has been tirelessly plopping the date into each room each weekday (and now, in crunch time, each day).

Then I watched a video that mentioned the Campfire interface to HUBOT, github's little slave server that handles all sorts of tasks. What if we could type a command into Campfire and have it actually do something?

But what? A first use case quickly suggested itself.

One of our Campfire rooms is devoted to server issues, and my boss and I often, while chatting in there, make a comment such as "todo: update deployment instructions."

How often do you think we actually remember those todos? Did you guess "at least rarely"? You may have overshot.

Some after-hours refactoring of the Ruby scripts I originally wrote plus a bit of tinkering with a new script, and I had a very simple command parser. Now, when you type "@SimCityBot todo blah blah blah," you'll get an email saying something like "You wanted to be reminded to blah blah blah." That doesn't guarantee the task will get done, of course, but it does make it  less ephemeral.

The script is pretty straightforward: it polls each room looking for messages that start with "@SimCityBot," and then invokes a method with the same name as the first word in the text. That means adding a new command is now a simple matter of adding a single method to a file. Yay for Ruby's metaprogramming support! The script also maintains a YAML file that keeps track of the most recent messages in each room. This ensures that when the script is restarted, it doesn't respond a second time to every command it sees.

I had to add support to our Campfire library for uploading images in order to post our "slowest calls" graphs each day. Once that work was done, adding an "image" command was a single call. Give it a URL, and the the script downloads that image and re-uploads it to whatever room the command appeared in.

Next up is a command to kick off Hudson builds. For that one, of course, I'll want to spin off a process that can monitor the build and report back when it's done. A co-worker suggested listing Emeryville food trucks. (Which I maintain as a Twitter list.)

There's lots of things that the script isn't good at. Handling a command blocks until it's done. Error reporting is minimal. It doesn't support multi-word commands. But it's a little trinket I can poke at and have fun with.

Is this the most important thing I could be doing for SimCity? Let's hope not. But at the end of a long day, sometimes I need a break, and my breaks from programming are … other programming projects! SimCityBot provides a refreshing distraction that often buoys my mood and gives me a nice close to the day. That's also part of why I've not just installed HUBOT. There's less fun in that.

But the catalog of all HUBOT scripts is an inspiring read. AWS status checks? Graphite graphs? The latest XKCD? When is break time again?

Scala and Java: A Simple Application

My boss recently asked me how I'd build out SimCity's online systems today if I knew everything three years ago that I know now. I didn't hesitate. "I'd take a good, long look at Scala and, by extension, Lift."

Scala has a lot that appeals to the me of today. I like its hand-in-hand support for functional programming and emphasis on immutable objects. I've gotten used to both concepts with Erlang, and I've come to appreciate that programming paradigm for building robust, scalable systems. But Scala also offers imperative syntax and mutable objects if you need them.

Scala has native support for the actor model abstraction of concurrency, which I first encountered with Erlang (Scala's syntax is openly lifted from that language's). The actor model makes it much easier to manage and reason about concurrent code, and Scala supports two major implementations: actors tied to particular threads or event-based actors thrown onto whatever thread is available, maximizing resource utilization.

And, unlike something like node.js or even Erlang, Scala has a huge universe of libraries at its disposal thanks to its bytecode-compliance with Java.

All good stuff. I thought it was time to do something real with it.

Before I dove into a large system, I thought I'd write a simple application. We have a group of binary files on SimCity that are very important to the game but, being binary, aren't easy to debug when something goes wrong. So I thought I'd do a quick project on my own to write a Java-based parser for the files and compare it to a Scala-based version. Little admin tools like this or other low-risk sections of the code base are often good ways to try out new tech and see if it will fit into the larger project. This code didn't leverage any of the concurrency systems in Scala; I just wanted a simple program.

One reason people like Scala — and, indeed, the many other JVM-compatible languages — is conciseness. Spend any time at all with Scala, Ruby, or — it sometimes seems — any other language, and Java's verboseness begins to feel like cement around your hands, sucking time and productivity away from your programming. Plus, more code, even Java's boilerplate, means more potential bugs.

As a simple measure, I compared the the non-whitespace characters in my two versions. I structured the programs the same way, mirroring class structures and refactored methods. But I used the support that each language gave me for keeping things concise.

The Java version was 5258 characters. The Scala version? 3099. The Java version was almost 70 percent larger.

Java Code	Scala Code
5258	3099

Scala's biggest single win was with a file full of small classes that defined types of data within the file I was parsing. The Java version was 160 percent bigger than the Scala one.

This makes sense. Let's say you wanted an immutable class in Java to represent a point in 3D space. This is about as concise as you can get it.

public class Point {
   public final float x, y, z;
   public Point(float _x, float _y, float _z) {
      x = _x;
      y = _y;
      z = _z;
   }
}

Here's the equivalent Scala code.

class Point(val x: Float, val y: Float, val z: Float)

But Scala offers lots of little aids as well. You rarely need Java's omnipresent semicolons; you don't need to declare types as often, since Scala can usually infer them; you don't need to explicitly type "return" at the end of a function, because the last result in the function is the return value; you don't have to declare that you throw exceptions. The list of little things goes on and obviously adds up.

Functional programming, too, offers some conciseness. I needed a routine to read an unsigned int out of a variable-length byte array. In the Java version, I wrote this:

private static int byteArrayToInt(byte[] bytes) throws IOException {
long retVal = 0;
    for (int i = 0; i < bytes.length; i++) {
        retVal = (retVal << 8) | ((long)bytes[i] & 0xff);
    }
    return (int)retVal;
}

In the Scala version, I wrote this:

private def byteArrayToInt(bytes: Array[Byte]) = {
    ((0L /: bytes) {(current,newByte) => (current << 8).toLong | (newByte & 0xff).toLong}).toInt
}

(The references to longs in this int-parsing code are to cope with the fact that I needed to read very large unsigned ints from the files, which Java defaults to interpreting as signed integers. The way to get around that is to write into a larger memory space, namely a long.)

You could argue that the Scala version is concise to the point of obtuseness, even if you're familiar with the functional-programming mainstay foldLeft operation it represents. I agree that there's a balance to be struck. In particular, I'm not sold on the /: operator for foldLeft; I might opt for spelling it out to be more clear.

For functional programming geeks, note that, to the extent it can, Scala offers tail-call optimization on recursive calls.

But things weren't all sunshine and roses on the Scala side. Here is the average time to run my program for each version, timed over 1000 iterations.

Java Time	Scala Time
88ms	182ms

To some extent, I expected this. Scala has to compile down to Java bytecode, which means that all that syntactic sugar and functional programming and closure support must turn into Java concepts somewhere. Even my little program generates a slew of extra classes and, presumably, lots of extra code that has to be navigated. Also, I think it's reasonable to imagine that immutable objects necessarily mean that new objects have to be created more often than they would in mutable space, where you can change an object directly. Finally, I've been working with Java in one form or another for 16 years or so; I've been working in Scala for about three days. So I'm likely missing out on performance tips.

Though I admit this seems like a huge difference for some extra classes and objects and missing an optimization step or two. Even if it's correct, I'm still of the mindset that greater productivity and easier, safer concurrency are big wins. (Note that you could always switch to imperative mode in key sections if performance demanded it, in much the same way that some sites offload work to C programs.)

If I were really honest about how I'd rebuild SimCity, I'd probably use Erlang, where you have to do things functionally, have a virtual machine that supports what you're doing, and have native systems for handling failures with aplomb. But Scala at least offers the potential of hiring from the pool of Java programmers, whereas Erlang really doesn't. (On the other hand, the vast majority of Java programmers I've seen seem to be couched safely and comfortably in Java, so wouldn't necessarily adapt. But Erlang would be a way bigger change, I think.)

I'm going to keep plunking away at Scala and try to build something a bit more real with it. Event-based actors might be a bit slower, but if they can scale vastly better, that may matter more to a site.

Copying On S3

The question recently arose: Is it faster to copy within buckets on S3 than it is to copy between buckets?

A quick script provided an answer. I copied a 100K file 100 times for each test and averaged the results (which are in seconds).


Avg. time to make copy between buckets: 0.10705331
Avg. time to make copy within bucket: 0.10522299

A second test produced similar results (very slightly slower in both cases).

And here's the Ruby script I threw together. It uses the aws-sdk gem.

# get buckets
s3 = AWS::S3.new
bucket1 = s3.buckets['dfsbucket1']
bucket2 = s3.buckets['dfsbucket2']

# get an object from bucket 1
random_file = bucket1.objects['191111308/state_file']

start = Time.now
copies = 100
(1 .. copies).each do |i|
  random_file.copy_to("test_file#{i}", {
     :bucket => bucket2
  })
end
puts "Avg. time to make copy between buckets: #{(Time.now - start)/copies}"

start = Time.now
(1..copies).each {|i| random_file.copy_to("test_file#{i}")}
puts "Avg. time to make copy within bucket: #{(Time.now - start)/copies}"