CommonStore Solution Deployment Consideration: Initial Offload and Large System Deployment
Published 29 December 2006
Authors: Wei-Dong Zhu
This technote is extracted from the IBM Redbooks publication, Best Practices for Setting Up an IBM CommonStore Solution, SG24-7525. It focuses on two areas of CommonStore solution deployment: initial offload and large system deployment. We discuss some of the considerations you should have to ensure successful deployment.
Initial offload consideration
When a CommonStore solution is introduced, it is almost always impossible to start with the normal day-to-day business of archiving e-mails. In almost all deployments, you have to deal with a huge amount of historical data stored in the mail files. Hence, an initial offload is often the first phase for which the CommonStore solution is used right after its deployment. The initial offload is a process to archive the historical e-mail from mail files. As already stated, this normally happens during the introduction phase of a CommonStore solution or if the archiving policy changes dramatically resulting in larger amounts of e-mails to be archived. For instance, a CommonStore solution is enabled with a policy to archive all documents older than one year. That leads to an initial offload of e-mail older than one year, regardless when the database was created. If a mail file is a couple of years old with a size of 1 GB and 60% of the documents are older than one year, then there is an initial offload of approximately 600 MB from that mail file. In contrast to that, during the daily archiving, only a handful of e-mails are expected to be archived per mail file after the historical data is archived already.
From our experience, there are two approaches for the initial offload: A time-driven setup or a content-driven setup. When using a time-driven setup, you normally start to archive the oldest e-mails first, and then adjust the archiving policy in reasonable steps towards the final setting that will be used for the daily archiving. The content-driven approach requires you to have much more knowledge about the content of the mail files than the time-driven approach to create the right archiving policies. In case you know that there is a particular amount of a certain type of document (for example, to be specified by Form name), these documents can be archived first, and then the archiving policy can be adjusted to match the next type of document. The content-driven approach can be also based on special content such as subject fields always containing a special value or the size of the e-mails.
Regardless of the approach you choose, it is advisable to accompany the initial offload by attentive monitoring of the data throughput. The initial offload should not influence the performance of the mail system under any circumstances in order to avoid dissatisfaction of the users.
Large system deployment consideration
Deploying an IT solution in a large, distributed enterprise environment is always a challenging task. A CommonStore solution is no different in such a situation. We distinguish between the two main areas of operation of a CommonStore solution: The archiving process and the retrieving/searching processes.
For the archiving process, answer the following questions before or during deployment:
- Where are the back-end components of the mail system located (that is, centralized or distributed)?
- Is the network capacity in place to support a distributed CommonStore solution?
- Is there a backup plan in place for the CommonStore solution? How will the recovery processes work in case of data loss?
- Can the CommonStore solution be monitored in a decent way?
- How are new mail systems integrated and can the CommonStore solution deployment deal with moving sources (for example, relocation of mail servers)?
For the retrieving/searching processes, consider the following questions:
- Is the deployed solution capable of serving thousands of user requests when using document retrieval via HTTP?
- Does the solution address the needs of remote or offline users in terms of the support process?
- Can data be retrieved or searched even if a user is relocated to a department in another location?
- Will requests that are created offline be processed after reconnection to the network?
- How difficult would it be to re-create the original condition of a mail file with all e-mails if necessary?
Answering these questions and taking them into consideration with specific concentration on automation, monitoring, and support will help you to deploy a successful CommonStore solution.
This material has not been submitted to any formal IBM test and is published AS IS. It has not been the subject of rigorous review. IBM assumes no responsibility for its accuracy or completeness. The use of this information or the implementation of any of these techniques is a client responsibility and depends upon the client's ability to evaluate and integrate them into the client's operational environment.
Follow IBM Redbooks
Follow IBM Redbooks