From owner-freebsd-isp@FreeBSD.ORG Tue Aug 5 15:11:57 2003 Return-Path: Delivered-To: freebsd-isp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 4775137B401 for ; Tue, 5 Aug 2003 15:11:57 -0700 (PDT) Received: from elk.hughes.com.au (elk.hughes.com.au [198.78.66.227]) by mx1.FreeBSD.org (Postfix) with ESMTP id AA83843FB1 for ; Tue, 5 Aug 2003 15:11:56 -0700 (PDT) (envelope-from bambi@Hughes.com.au) Received: from elk.hughes.com.au (elk [198.78.66.227]) by elk.hughes.com.au (8.12.6/8.12.3) with ESMTP id h75MBvWN036059; Tue, 5 Aug 2003 15:11:57 -0700 (PDT) (envelope-from bambi@Hughes.com.au) Received: from localhost (bambi@localhost)h75MBvqm036052; Tue, 5 Aug 2003 15:11:57 -0700 (PDT) X-Authentication-Warning: elk.hughes.com.au: bambi owned process doing -bs Date: Tue, 5 Aug 2003 15:11:56 -0700 (PDT) From: "David J. Hughes" To: Scott Lambert In-Reply-To: <20030805135045.GA672@laptop.lambertfam.org> Message-ID: <20030805150102.N25983@elk.hughes.com.au> References: <20030804182222.A5C055553E@titanic.medinet.si> <20030805073757.G78439@titanic.medinet.si> <20030805135045.GA672@laptop.lambertfam.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII cc: FreeBSD-ISP@freebsd.org Subject: Re: DNS Server Farm X-BeenThere: freebsd-isp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Internet Services Providers List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 05 Aug 2003 22:11:57 -0000 On Tue, 5 Aug 2003, Scott Lambert wrote: > If you want a load balancer, and they are cheap, why not put one in each > regional hub and have several smaller clusters? Because if you have 50 regional POPs then the management of either 50 or 100 DNS servers (2 per POP to ensure service continuity) becomes an issue. As does the cost. Remember the original poster was talking about a network with > 1,000,000 subscribers so there'll probably be quite a few POPs. If there is no "sacred heart" of your network that is resilient to multiple single failures, then build 2 pairs of load balanced servers and place them at strategic locations in your core. That's 4 servers and 4 load balancers (running HA mode) and not a single point of failure in sight if it's done right. With ServerIrons going for sub $1,000 on ebay it's far cheaper and a hell of a lot easier to manage than a huge anycast solution. IMHO naturally :) Bambi ...