What is the current status and roadmap for FreeRadius?

Alan DeKok aland at deployingradius.com
Sat Mar 16 14:55:16 CET 2019

On Mar 15, 2019, at 9:40 PM, Arran Cudbard-Bell <a.cudbardb at freeradius.org> wrote:
>> What we are really feeling over here is a lack of modern behavior and
>> features.
> As someone who's done multiple, complex, commercial deployments with
> FreeRADIUS I disagree.

  As an aside, C++ and Antrl4 are tools, not features.  Using "cool tools" just because they're "cool" is the sign of cargo cult programming.

  If you want to look at cargo cult programming, look no further than Kea, the "new" ISC DHCP server.

  The old ISC DHCP server was written in C, using (quite frankly) terrible engineering practices.  Those practices turned out to be problematic, so they went to "cool tools" like C++, Javascript, libboost, etc.  But they didn't change their poor engineering.

  The result is a DHCP server where the config files are javascript.  Which makes them difficult to edit by mere mortals.  And where ISC DHCP had *zero* database integration, Kea has MySQL and PostgreSQL.  That sounds good, but...

  FreeRADIUS has support for many, many, more databases than that.  FreeRADIUS can put leases into *any* SQL database.  Or Redis.  And, in v4, we get 2-5x the performance of Kea for DHCP leases.  With a *real* policy language, that doesn't require people to write C++ plugins to do anything complex.

  So yes, FreeRADIUS is a *better* and *more configurable* and a *faster* DHCP server than one written using "cool tools".

  I have no idea what "modern behaviour and features" means.  And I doubt that the original poster does, either.  If he had, he would have explained instead of posting inflammatory and content-free comments.

  Alan DeKok.

More information about the Freeradius-Users mailing list