1 Copper

Is there any tool for sizing RepliStor?

I want to know How  big machine I need to deploy to support up to 450 remote Data replication to one central localization.

Labels (1)
0 Kudos
2 Replies
3 Argentium

Re: Is there any tool for sizing RepliStor?

There is no sizing tool that exists for RepliStor.

There is a RepliStor Utility that will alolow you to install/upgrade machines in a large environment.  It is available for download via PowerLink.

The main consideration for you will be the target maching.  If this is a many-to-one replication topology, you will want to ensure that the Target has sufficient processing and network availability to manage all of the incoming updates from the applicable Sources,


2 Iron

Re: Is there any tool for sizing RepliStor?


There is no separate tool for sizing a RepliStor Many-to-One configuration.  The closest thing is to install RepliStor on One Source node and then on the target node.  There are built-in reports that will provide statistical and graphical information about how busy the Target node is and the number of Source nodes it can handle can be roughly extrapolated based on this information.

However, based on my experience, 450 to one is probably not a realistic expectation.  I have seen customers successfully replicate 50-60 machines to a single node where there was not a high rate of data change.  I have also seen customers who cannot exceed about 10 to one.

There are simply too many variables involved to provide a single answer to all customers.  You need to actor in things like the following: Connection Speed and Latency, disk read and write throughput and caching, rate of change on each Source node, number of files being replicated per folder, Active Directory access on both nodes and many other factors.

In general, since RS is a host based replication S/W, each Source node is a variable as well as the connections between the Sources and Targets.  It also depends greatly on what types of files are being replicate.  Database files (any binary file) compresses less than a text file.  Since RS uses data compression to speed the transfer, those are critical factors as well.

Sorry I can't offer more specific information but each implementation must be taken as a unique piece to a larger puzzle.  There just isn't one formula that works in every instance.  Thanks.