Establishing patent infringement can be difficult at the best of times but when the technology alleged to infringe a patent is primarily software, there can be extra hurdles for the patent owner. Some recent court decisions reveal some of those challenges.
To prove infringement, a patentee has to show that the activities of the defendant fall within the scope of the claims of the asserted patent. Photos, and engineering drawings or even samples of the product can be used to prove the features of the product or process at issue.
When the patent relates to a particular a software implementation or software feature, it may be necessary to look at the software source code, as well as the execution of the software in operation to understand and establish how it works.
In a recent patent infringement trial decision from Justice Zinn (Google LLC v. Sonos, Inc., 2022 FC 1116) focusing on an audio processing system, there were several issues he had to consider relating to the software implementation of the defendant’s product.
First, one of the plaintiff’s experts had relied on excerpts of the software code extracted by a second expert. When the second expert witness was not called at trial, the trial judge had to weigh the evidence of between one expert who had only reviewed extracts of the software and the defendant’s expert who had reviewed the entirety of the source code. Justice Zinn held that where there was differences, he favoured the expert who had reviewed the entirety of the source code.
A second issue relating to the software was whether the defendant could directly infringe the patent. The Court found that the defendant’s product “out of the box” did not have the configured covered by the patent and therefore there could not be direct infringement by the defendant. Instead, the configuration at issue was only used after the end-user enabled a voice assistant feature. Because the defendant provided the device, and the option to the user to enable the feature, the Court would have found liability for induced infringement if the feature, as enabled by the user, contained all the features. Ultimately, the Court concluded there was no liability because the device’s features as enabled, did not change the order of certain process steps as required by the asserted claim. This decision points to the importance of clearly identifying which features, and the configuration of those features, that are alleged to infringe and in what circumstances those configured features are actually used.
In another recent decision, Farmobile, LLC v. Farmers Edge Inc., 2022 FC 22, in a pre-trial procedural decision, the defendant sought to bifurcate aspects of the scheduled trial, or alternatively adjourn the trial, to address new iterations of its software. Earlier in the proceeding, the defendant had announced a week prior to the scheduled trial that it had changed its software to implement what it claimed to be an alternative which its expert had described as being non-infringing. With the trial adjourned to a new date as a result of these developments, further expert reports were exchanged about whether the changed version was still infringing. In addition, the defendant implemented a further software update, which it said did not infringe under either expert’s construction of the patent.
The defendant sought to bifurcate the trial to only address the earlier versions of its software at a first trial, with the updated version of the software dealt with later, or in the alternative, an adjournment of the entire trial. Further discovery and expert reports were required as a result of the second software update.
The Court rejected the request to further postpone the trial saying that the parties likely had enough time to prepare expert reports before the trial. The Court also recognized that the defendant may implement a further software update but said it would deal with that later, if it occurred, stating:
I am certainly sympathetic to [the defendant’s] assertion that it must be permitted to amend its software to avoid infringement. Indeed, it would seem odd for [the plaintiff] to insist that [the defendant] not change its software, thereby continuing to infringe (on [the plaintiff’s] allegation), so that [the plaintiff] can obtain at trial an injunction requiring [the Defendant] not to infringe. I am also sympathetic to [the Plaintiff’s] argument that at some point the issues for trial must be fixed so that the parties can proceed to trial in an orderly fashion.
With software being often updated, to include new features, fix bugs or address security vulnerabilities, it can be difficult for parties to identify which features, or implementation of features, are at issue for a specific time period. With many systems being implemented “in the cloud”, a software update can immediately change the behaviour for the entire system, even for previously sold consumer devices. This can be particularly important where a defendant purports to implement a non-infringing alternative to avoid potential injunctive relief. As recently found in a case I was involved in, Rovi Guides, Inc. v. Videotron Ltd., 2022 FC 874, the cost to implement a non-infringing alternative can an upper bound on the financial relief obtainable by a plaintiff.
Because of the complexity of software configurations, including on distributed and cloud-based platforms and the incorporation of modifications to software, proving infringement of software implemented systems can raise additional issues for plaintiffs in patent infringement proceedings.
This column is being published on the 10 year anniversary of my first column for Slaw. Thanks to Simon Fodden in the early days and now Steve Matthews for this opportunity to share my thoughts on intellectual property and contribute to the discussions on Slaw.