From fork-admin@xent.com Tue Oct 8 10:56:42 2002 Return-Path: Delivered-To: yyyy@localhost.example.com Received: from localhost (jalapeno [127.0.0.1]) by jmason.org (Postfix) with ESMTP id CCC2316F16 for ; Tue, 8 Oct 2002 10:56:41 +0100 (IST) Received: from jalapeno [127.0.0.1] by localhost with IMAP (fetchmail-5.9.0) for jm@localhost (single-drop); Tue, 08 Oct 2002 10:56:41 +0100 (IST) Received: from xent.com ([64.161.22.236]) by dogma.slashnull.org (8.11.6/8.11.6) with ESMTP id g989KUK08679 for ; Tue, 8 Oct 2002 10:20:31 +0100 Received: from lair.xent.com (localhost [127.0.0.1]) by xent.com (Postfix) with ESMTP id 431202940AE; Tue, 8 Oct 2002 02:20:04 -0700 (PDT) Delivered-To: fork@example.com Received: from gremlin.ics.uci.edu (gremlin.ics.uci.edu [128.195.1.70]) by xent.com (Postfix) with ESMTP id D50442940AE; Tue, 8 Oct 2002 02:19:38 -0700 (PDT) Received: from localhost (godzilla.ics.uci.edu [128.195.1.58]) by gremlin.ics.uci.edu (8.12.5/8.12.5) with ESMTP id g989JCgB000573; Tue, 8 Oct 2002 02:19:12 -0700 (PDT) Subject: why is decentralization worth worrying about? Content-Type: text/plain; charset=WINDOWS-1252; format=flowed MIME-Version: 1.0 (Apple Message framework v482) From: Rohit Khare To: fork@example.com Message-Id: <2583F1FA-DA52-11D6-B1B1-000393A46DEA@alumni.caltech.edu> X-Mailer: Apple Mail (2.482) X-Mailscanner: Found to be clean Sender: fork-admin@xent.com Errors-To: fork-admin@xent.com X-Beenthere: fork@example.com X-Mailman-Version: 2.0.11 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: Friends of Rohit Khare List-Unsubscribe: , List-Archive: Date: Mon, 7 Oct 2002 17:09:00 -0700 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by dogma.slashnull.org id g989KUK08679 X-Spam-Status: No, hits=-5.2 required=5.0 tests=AWL,DATE_IN_PAST_06_12,KNOWN_MAILING_LIST, T_NONSENSE_FROM_30_40,USER_AGENT_APPLEMAIL version=2.50-cvs X-Spam-Level: Why am I so passionate about decentralization? Because I believe some of today’s most profound problems with networked applications are caused by centralization. Generically, a centralized political or economic system permits only one answer to a question, while decentralization permits many separate agents to hold different opinions of the same matter. In the specific context of software, centralized variables can only contain one valid value at a time. That limits us to only representing information A) according to the beliefs of a single agency, and B) that changes more slowly than it takes to propagate. Nevertheless, centralization is the basis for today’s most popular architectural style for developing network applications: client-server interaction using request-response communication protocols. I believe these are profound limitations, which we are already encountering in practice. Spam, for example, is in the eye of the beholder, yet our email protocols and tools do not acknowledge the separate interests of senders and receivers. Slamming, for another, unfairly advantages the bidder with the lowest-latency connection to a centralized auction server. Sharing ad-hoc wireless networks is yet a third example of decentralized resource allocation. Furthermore, as abstract as centralization-induced failures might seem today, these limits will _not_ improve as the cost of computing, storage, and communication bandwidth continue to plummet. Instead, the speed of light and human independence constitute _fundamental_ limits to centralized information representation, and hence centralized software architecture.