15

I've been doing socket programming for a while in C++, and kind of got tired of having to write the same code to handle for errors, serializing / deserializing data, etc.

Are there programming languages out there that have first-class support for distributed system?

sivabudh
  • 31,807
  • 63
  • 162
  • 228

5 Answers5

22

Erlang, as described by Wikipedia:

It was designed by Ericsson to support distributed, fault-tolerant, soft-real-time, non-stop applications.

You might also want to read the Distributed Erlang section of their manual.

However, note that Erlang is a functional language and will require a much different paradigm of thought as compared to C++.

Mark Rushakoff
  • 249,864
  • 45
  • 407
  • 398
  • 3
    Erlang is definitely the poster-child. While there's literally thousands of languages that are specifically designed for distributed programming, Erlang is by far the most mainstream (and in fact pretty much the only one). – Jörg W Mittag Jul 23 '10 at 23:26
  • I chose to accept this answer because Erlang is used in real-world systems and is mainstream, even though, as pointed out by Jorg, there are thousand others built for distributed system. – sivabudh Jul 23 '10 at 23:45
  • I dont think Scala "natively" supports distributed programming. However, with Terracotta, see: http://jonasboner.com/2008/01/25/clustering-scala-actors-with-terracotta.html. Or see Swarm: http://lambda-the-ultimate.org/node/3626 – sivabudh Jul 24 '10 at 00:33
  • 1
    [elixir](https://stackoverflow.com/questions/37964709/are-there-things-elixir-can-do-that-erlang-cannot-or-vice-versa) is a "simpler" version of erlang – milahu Jan 26 '22 at 10:45
  • @MilaNautikus how? – Nicolás Gómez May 27 '22 at 21:56
17

Go-Lang from Google is a pretty new language. It seems that among its many attributes, it may some day be suitable for large distributed systems requiring a lot of message queues to achieve scalable consistent and reliable behaviours, at least according to these folks at heroku.

Go seems to be focused on concurrency issues, threading primitives in the language, and so on, and this is perhaps a necessary-but-not-quite-sufficient starting point for distributed systems. Perhaps their thoughts will be helpful to you. I wouldn't call Go-lang's support for distributed systems "first-class", but rather, say that it would be possible to build a first class distributed-systems framework using Go's library and language primitives.

Update: I'm less impressed with Go several years later. I think it suffers from some sad and limited thinking on the part of its authors. I think its decisions on fault and exception handling are retrograde, and render the language unusable.

Update-2016: I'm actually impressed with Go again. I now think in terms of large team development where having N-factorial implementation options leads to N-factorial different coding tarpits. At least Go doesn't seem to have labrea scale tarpits, only certain conventional mudwallows. They absolutely love tabs and will insert them into your code for you if you fail to love them enough.

Warren P
  • 65,725
  • 40
  • 181
  • 316
  • 6
    The Updates on your thinking about Go over such a long time span along with the reasons are most interesting. Thank you. – masterxilo Oct 30 '17 at 11:27
  • go isn't distributed friendly as erlang is when scaling within regions level large size. either parallel or concurrent, those feature are differ from distributed requirement. – Oshan Wisumperuma Jul 10 '19 at 08:44
  • @Warren P What are your thoughts on Go simplicity? Stated many times by Rob Pike. – Nicolás Gómez May 27 '22 at 21:58
  • It's not for everyone, Go is the perfect language for Rob Pike or people who are forced to work with Rob Pike. – Warren P Jun 03 '22 at 17:26
  • I think the updates over a long period are interesting. I have come to go lately and it feels like I can finally learn Oberon and Occam which I never got around to earlier in my career. I am still looking for reliable distributed channels solution. I think [NSQ](https://nsq.io/overview/features_and_guarantees.html) is the solution and need to play with it. – hum3 Feb 18 '23 at 12:31
  • By the way I stopped using Stack Overflow regularly when they switched to closing all questions that call for any kind of discussion, or offering an opinion, or making a choice where there was not one right answer, such as this question. StackOverflow died on the day they decided to kill it this way. In 2023, StackOverflow is where you go to find things that are already in every reference manual on earth. – Warren P Feb 20 '23 at 16:58
4

Bloom is a new domain-specific language for distributed programming. The current alpha release is embedded in Ruby, and targeted at early adopters. Bloom leverages new research on "CALM" analysis to provide tools that pinpoint distributed consistency and coordination issues in your code.

Warren P
  • 65,725
  • 40
  • 181
  • 316
jmh
  • 41
  • 1
1

Reia is a scripting language for distributed system:

Reia aims to expose the powerful capabilities of Erlang in a way which is easier for your average programmer to understand. It aims to bring the beauty and simplicity of Ruby, a language which is easy and fun to program in, to Erlang, a language which very few will think of as easy or fun to use.

sivabudh
  • 31,807
  • 63
  • 162
  • 228
1

Parallel Python is a python module which provides mechanism for parallel execution of python code on SMP (systems with multiple processors or cores) and clusters (computers connected via network):

Features:

  • Parallel execution of python code on SMP and clusters
  • Easy to understand and implement job-based parallelization technique (easy to convert serial application in parallel)
  • Automatic detection of the optimal configuration (by default the number of worker processes is set to the number of effective processors)
  • Dynamic processors allocation (number of worker processes can be changed at runtime)
  • Low overhead for subsequent jobs with the same function (transparent caching is implemented to decrease the overhead)
  • Dynamic load balancing (jobs are distributed between processors at runtime)
  • Fault-tolerance (if one of the nodes fails tasks are rescheduled on others)
  • Auto-discovery of computational resources
  • Dynamic allocation of computational resources (consequence of auto-discovery and fault-tolerance)
  • SHA based authentication for network connections
  • Cross-platform portability and interoperability (Windows, Linux, Unix, Mac OS X)
  • Cross-architecture portability and interoperability (x86, x86-64, etc.)
  • Open source

One can get a quick idea of how the code might look by looking at the quick-start guide for clusters.

Caleb Hattingh
  • 9,005
  • 2
  • 31
  • 44