Return-Path: Delivered-To: apmail-incubator-cassandra-dev-archive@minotaur.apache.org Received: (qmail 53189 invoked from network); 2 Apr 2009 17:09:33 -0000 Received: from hermes.apache.org (HELO mail.apache.org) (140.211.11.3) by minotaur.apache.org with SMTP; 2 Apr 2009 17:09:33 -0000 Received: (qmail 39229 invoked by uid 500); 2 Apr 2009 17:09:33 -0000 Delivered-To: apmail-incubator-cassandra-dev-archive@incubator.apache.org Received: (qmail 39184 invoked by uid 500); 2 Apr 2009 17:09:33 -0000 Mailing-List: contact cassandra-dev-help@incubator.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: cassandra-dev@incubator.apache.org Delivered-To: mailing list cassandra-dev@incubator.apache.org Received: (qmail 39174 invoked by uid 99); 2 Apr 2009 17:09:33 -0000 Received: from athena.apache.org (HELO athena.apache.org) (140.211.11.136) by apache.org (qpsmtpd/0.29) with ESMTP; Thu, 02 Apr 2009 17:09:33 +0000 X-ASF-Spam-Status: No, hits=-1.8 required=10.0 tests=HTML_MESSAGE,RCVD_IN_DNSWL_MED X-Spam-Check-By: apache.org Received-SPF: unknown (athena.apache.org: error in processing during lookup of junrao@almaden.ibm.com) Received: from [32.97.182.146] (HELO e6.ny.us.ibm.com) (32.97.182.146) by apache.org (qpsmtpd/0.29) with ESMTP; Thu, 02 Apr 2009 17:09:24 +0000 Received: from d01relay02.pok.ibm.com (d01relay02.pok.ibm.com [9.56.227.234]) by e6.ny.us.ibm.com (8.13.1/8.13.1) with ESMTP id n32HAbZA026897 for ; Thu, 2 Apr 2009 13:10:37 -0400 Received: from d01av05.pok.ibm.com (d01av05.pok.ibm.com [9.56.224.195]) by d01relay02.pok.ibm.com (8.13.8/8.13.8/NCO v9.2) with ESMTP id n32H93ja192276 for ; Thu, 2 Apr 2009 13:09:03 -0400 Received: from d01av05.pok.ibm.com (loopback [127.0.0.1]) by d01av05.pok.ibm.com (8.13.1/8.13.3) with ESMTP id n32H93DB025744 for ; Thu, 2 Apr 2009 13:09:03 -0400 Received: from d01ml604.pok.ibm.com (d01ml604.pok.ibm.com [9.56.227.90]) by d01av05.pok.ibm.com (8.13.1/8.12.11) with ESMTP id n32H93Cu025741 for ; Thu, 2 Apr 2009 13:09:03 -0400 In-Reply-To: Subject: Re: OPHF To: cassandra-dev@incubator.apache.org X-Mailer: Lotus Notes Release 7.0 HF277 June 21, 2006 Message-ID: From: Jun Rao Date: Thu, 2 Apr 2009 10:09:04 -0700 X-MIMETrack: Serialize by Router on D01ML604/01/M/IBM(Release 8.5|December 05, 2008) at 04/02/2009 13:09:02 MIME-Version: 1.0 Content-type: multipart/alternative; Boundary="0__=07BBFF1FDFCEAE218f9e8a93df938690918c07BBFF1FDFCEAE21" Content-Disposition: inline X-Virus-Checked: Checked by ClamAV on apache.org --0__=07BBFF1FDFCEAE218f9e8a93df938690918c07BBFF1FDFCEAE21 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: quoted-printable If the partitioner maps String->token, then the load balancer can be oblivious to the partitioner and do the balancing by reassigning tokens= to EndPoint. On the other hand, if the partitioner maps String->EndPoint, = then the load balancer has to tell the partitioner that it has changed some = of the decision the partitioner already made. How will this work? Jun IBM Almaden Research Center K55/B1, 650 Harry Road, San Jose, CA 95120-6099 junrao@almaden.ibm.com Jonathan Ellis wrote on 04/02/2009 09:51:26 AM: > > On Thu, Apr 2, 2009 at 10:38 AM, Jun Rao wro= te: > > > > Johnathan, > > > > Could you explain a bit more how the pluggable partitioner (that provides > > String-> EndPoint mapping) works? > > > > In particular, > > 1. How does the partitioner obtain the list of EndPoints in the sys= tem? > > Same way it does now, through Gossip. The main change is from > > private Map tokenToEndPointMap_ > > to > > private Map tokenToEndPointMap_ > > where Token can either be implemented as a wrapper around BigIntegero= r String. > > > 2. What happens when EndPoints are added/removed from the system? D= oes the > > partitioner have to remember the history of the EndPoints? > > Ditto. > > > 3. How does the partitioner coordinate with the load-balancer? Supp= ose some > > rows are moved from one EndPoint to another. How is that informatio= n > > communicated and used by the partitioner so that routing remains correct? > > Load balancing isn't complete yet; I didn't mean to imply otherwise. > So I don't know for sure. Avinash has probably thought about this > more. But having Token be abstract shouldn't affect the algorithm. > > -Jonathan= --0__=07BBFF1FDFCEAE218f9e8a93df938690918c07BBFF1FDFCEAE21--