3 Infrastructure Management Pains & How to Prevent Them


Working behind the scenes to support case teams, IT and infrastructure experts handle frequent requests for reports on and improvements to their organization’s e-discovery environment. When case sizes explode without warning, it can be tough to keep up.

Since 2010, we’ve seen a 400 percent growth in the 100 largest cases in Relativity, and it’s easy to see how. Say you recently helped stage an environment for a case with an expected 100,000 documents. Four weeks later, new PSTs are discovered on an older file share, and the case has grown to 1,000,000 documents. The case team needs help pivoting to handle the new data. Two months after that, the same thing happens on another case—and year over year, you find your average case size is growing significantly. 

If this is happening to you, it’s time to evaluate your infrastructure’s maturing needs. Here’s a look at three growing pains you might face along the way—and how you can avoid them.

Growing Pain #1: You need to maintain performance for unexpectedly large—and growing—case sizes.

The case team is working on cases that are many times bigger than their initially projected size and seeing performance lagging. They’ve just processed new data, and searching and doc-to-doc viewing are slowing down. You need to jump in and allocate the right resources to meet the new requirements—a process that can mean a lot of late nights and weekends moving cases around and downtime for the review environment.

Recommendation: Use a distributed NoSQL storage option that scales horizontally.

Find out what remedies exist before vertical scaling becomes a consistent problem in your environment. Work closely with your case team to understand how the software they’re using scales as case sizes or volumes increase. How can you prepare your environment ahead of reaching an infrastructure limit? Is there a NoSQL or similar option available? Touch base frequently to ensure things are working as they should, and plan changes to infrastructure where they’re warranted.

Distributed NoSQL tools automatically balance storage and compute power across the resources you already have supporting the environment, with minimal administrative effort and no performance interruptions—and more data distribution is automated as your cases grow. Adding new hardware takes significantly less effort. If you’re running Relativity, Data Grid offers this NoSQL approach for the platform.

Growing Pain #2: Reviewer mistakes multiply with data growth and take a lot of manual investigation to identify and resolve.

In an environment with so many documents, one reviewer’s simple mistake—like mass editing the wrong list of documents or misunderstanding a fact about the case for an afternoon—can have a big impact. That means you’ll spend a lot of time running complex search queries and scripts on audit data to help the case team identify the documents the reviewer coded and resolve them in enough time to meet their production deadline.

Recommendation: Get familiar with your software’s audit search and reporting options. 

In situations like these, half your frustration is spent working around the case team’s schedule. If you simply can’t run some of those searches without impacting environment performance and interfering with ongoing work, you’ll need to run them after hours or come up with a plan to get a copy of the data.


It’s possible, however, that your organization’s platform offers searching, reporting, and data visualization tools to supplement or even replace many of those SQL queries you’re used to running. Data Grid’s audit UI allows you to visually explore and analyze your environment’s audit data in real time without impacting review performance. With built-in dashboards, you can quickly pull the information the case team needs and share it with them. 

Growing Pain #3: Frequent back-and-forth with litigation support teams is slowing you down.

This one’s a compound ache caused by—and contributing to—both of the previous pains. As more data is added to your environment and larger projects become the norm, your case team will be coming to you more frequently to address performance problems and run queries. If requests are primarily reactive to problems with cases and they’re happening too often, they’ll slow down the case team’s progress on their projects and prevent you from taking care of other tasks.

Recommendation: Pursue an active partnership with litigation support and break your transactional habits.

Instead of lobbing performance upgrade requests and reports over silo walls, work closely with your legal and case teams on an ongoing basis. Analyzing your case size growth over time is a great way to start. Review your case load, document counts, and infrastructure needs over the past year or two. Compare that with your plans for the next year, and make sure you are in sync about your forecasts. 

If you’re implementing the first two recommendations, it’ll also be easier to provide case teams with more visibility into their environment’s health and management. Audit reporting built on NoSQL data stores without the need for back-end queries might even mean you can teach the case team to run reports on their own. You can save time and stay more plugged into their needs and priorities, improve collaboration on important data management initiatives, and get information to the right people more quickly.

Crystal Ye is a product manager at kCura, where she focuses on production support for Relativity and helps guide ongoing improvements to the platform.