1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
|
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<TITLE> [Nel] Hello and a Question</TITLE>
<LINK REL="Index" HREF="index.html" >
<LINK REL="made" HREF="mailto:lecroart%40nevrax.com">
<LINK REL="Previous" HREF="000421.html">
<LINK REL="Next" HREF="000422.html">
</HEAD>
<BODY BGCOLOR="#ffffff">
<H1>[Nel] Hello and a Question</H1>
<B>Vianney Lecroart</B>
<A HREF="mailto:lecroart%40nevrax.com"
TITLE="[Nel] Hello and a Question">lecroart@nevrax.com</A><BR>
<I>Fri, 15 Jun 2001 19:35:16 +0200</I>
<P><UL>
<LI> Previous message: <A HREF="000421.html">[Nel] Hello and a Question</A></li>
<LI> Next message: <A HREF="000422.html">[Nel] A bit of news</A></li>
<LI> <B>Messages sorted by:</B>
<a href="date.html#423">[ date ]</a>
<a href="thread.html#423">[ thread ]</a>
<a href="subject.html#423">[ subject ]</a>
<a href="author.html#423">[ author ]</a>
</LI>
</UL>
<HR>
<!--beginarticle-->
<PRE>><i> Anyway, my question is, why did you develop your own distributed
</I>><i> architecture? Why not adapt/extend/use an existing solution such as
</I>><i> CORBA or DCOM (urgh, I know, win32 only). CORBA has a fair amount of
</I>><i> support and it can be coaxed to have the characteristics you need. I am
</I>><i> not saying you should change, I was just wonder what led you to make
</I>><i> your own? It just seems that the overheads of designing a naming and
</I>><i> authentication service are something that CORBA would have done away with.
</I>
As you point out, DCOM isn't portable so we ruled it out :)
As for CORBA: Having worked a fair bit with it while at college, I concluded
that it lacked some of the functionality that we wanted. For example, we can
have several instances of the same service running simultaneously on our
shard and the naming service can either return us a list of all of them or
select the least loaded one (which CORBA cannot do). What's more, we found
that CORBA was excessively heavy for our requirements, which only require a
corner of the system. Our naming service only took a couple of days to
develop and remains entirely open to expansion, adaptation or tuning to keep
up with NeL's evolving needs.
We looked in depth into the ACE and TAO library (Real-time CORBA with TAO
(The ACE ORB)). The webpage seems down www.cs.wustl.edu/~schmidt/TAO.html,
but the lib is absolutely huge... one hour at least of compilation. We
didn't want to be dependent on such a large library just for a naming
service and a communication protocol.
Regards,
Vianney
</pre>
<!--endarticle-->
<HR>
<P><UL>
<!--threads-->
<LI> Previous message: <A HREF="000421.html">[Nel] Hello and a Question</A></li>
<LI> Next message: <A HREF="000422.html">[Nel] A bit of news</A></li>
<LI> <B>Messages sorted by:</B>
<a href="date.html#423">[ date ]</a>
<a href="thread.html#423">[ thread ]</a>
<a href="subject.html#423">[ subject ]</a>
<a href="author.html#423">[ author ]</a>
</LI>
</UL>
</body></html>
|
