Question on Enterprise Replication

Hello all, I have a question regarding Informix Enterprise Replication (ER), also known as CDR. We have several instances involved in a replication scheme right now and are looking to add a lot more to it. The scheme is a little complicated but I will try to simplify: Instance A is on Box 1 and is running IDS 7.31 on Sun and is the sync server Instance B is on Box 2 and is running IDS 7.31 on Sun Instance C is on Box 2 and is running IDS 7.31 on Sun Instances D - I are on Boxes 3-5 and are running IDS 9.21 on HP 11 (64-bit) The current replication scheme is: Eight tables are replicated from A to B Eight other tables are replicated from B to A All tables in A are replicated to C Two tables in A are replicated to each D through I We are planning on bringing eleven additional instances (J - T) into the mix - all running IDS 9.21 on HP 11 (64-bit). The seventeen IDS 9.21 instances are spread across six boxes, all HP 11 (64-bit). We will be replicating 211 tables from each Instance D through S to instance T. Instances A, B, and D through S are OLTP instances. Instances C and T are reporting instances. We are concerned about impacting Instance A. It is the sync server so it is not responsible for (nor involved with) moving the replicated data - it is only responsible for maintaining the authoritative global catalog. But we want to make sure that it will not be impacted when we add these additional instances. Instance A supports a critical application. Does anyone have experience with a sync server being impacted by the amount of data in the replication scheme? One additional question: We have experienced problems in the past when we are implementing an application change that involves some sort of mass update or load to replicated tables. Replication tends to choke when you try to run toomuch data through it too quickly. We now remove the affected replicates from their group and delete the replicates. After we make our change, we put everything back again. Are there any configuration parameters we can adjust to make replication more robust in handling mass updates and/or loads? Thank you in advance for any information. Rob Schmitz Rob.B.Schmitz@mail.sprint.com