Erlang and unsubstantiated reliability terms

Everybody and his dog babbles about Erlang being able to reach „nine nines reliability“ and claiming Jon Amstrong showed that in his Phd Thesis. Did he? No, not at all. To cite him:

„Evidence for the long-term operational stability of the system had also not been collected in any systematic way. For the Ericsson AXD301 the only information on the long-term stability of the system came from a power-point presentation showing some figures claiming that a major customer had run an 11 node system with a 99.9999999% reliability, though how these figure had been obtained was not documented.“ (Amstrong, „Making reliable distributed systems in the presence of sofware errors“, p 191, emphasis added by me)

So Erlang advocates: if you are claiming Erlang has prroven to reach „nine nienes“ of reliability you are dishonest or don’t know what you are talking about – choose what ever you prefer.

2 comments on “Erlang and unsubstantiated reliability terms

  1. alexis
    2008-10-16 at 00:07 #

    That’s a bit harsh

    That’s a bit harsh

    It would be correct to say that IF someone claims that Joe Armstrong’s thesis *proves* that Erlang has ‚reached nine nines‘ of reliability, THEN they have not read his thesis.

    But that is not the claim people tend to make!

    (Sigh – yes there is always too much hype in this industry).

    Just because Joe Armstrong does not show the nine nines claim in his PhD thesis, it does not then follow then that the claim has not been made by commercial users of Erlang, nor that it is unreasonable (or „dishonest“) to cite such claims.

    Joe Armstrong himself talks about this reliability number in favour of Erlang, and I would not accuse him of being dishonest or not knowing what he is talking about.

    I think people usually cite Joe’s 2002 talk „Concurrency Oriented programming in Erlang“ given at the Lightweight Languages Workshop 2002 (LL2) held at MIT. You can get the PDF from Joe’s site: .. see p.27 onwards.

    I think the fact that Joe cites this claim himself, in such a talk at such a setting, means that he does take it seriously and believe it even though „how these figure had been obtained has not been documented“.

    After all, his thesis is essentially an argument about how to make systems reliable and it uses AXD301 as a case study frequently („thought to be one of the most reliable products ever made by Ericsson“,

    I am not sure where you got this word ‚proven‘ from. Perhaps you mean that it is a word people throw around carelessly. Joe is not the source of all info about Erlang… For example, general reports of ejabberd tend to be positive — although admittedly often anecodotal.



    This comment was originally posted on 20070828T11:09:39

  2. mdornseif
    2008-10-16 at 00:07 #

    If you call it „rigorous“ instead of „harsh“ I’m perfectly happy with that description.

    I have a somewhat longer explanation of why I consider the „nine nines“ claim a disservice to the erlang community here.

    This comment was originally posted on 20070828T20:38:59

Kommentar verfassen

Trage deine Daten unten ein oder klicke ein Icon um dich einzuloggen:

Du kommentierst mit Deinem Abmelden / Ändern )


Du kommentierst mit Deinem Twitter-Konto. Abmelden / Ändern )


Du kommentierst mit Deinem Facebook-Konto. Abmelden / Ändern )

Google+ Foto

Du kommentierst mit Deinem Google+-Konto. Abmelden / Ändern )

Verbinde mit %s