Friday was the fifth day of trial in the patent infringement suit filed by Juniper Networks (NYSE:JNPR) against Palo Alto Networks (NYSE:PANW) in December 2011. For some background on the case, see my previous article discussing the judge's claim construction and summary judgment decision and my reviews of days one, two, three and four of the trial. I attended the trial Friday and provide my observations and thoughts here.
Before the jury was brought in to the courtroom for the trial to resume, the parties and judge discussed some of the instructions that she was thinking of reading to the jury at the end of the trial. Palo Alto objected to an instruction that the claims of Juniper's patents are not to be limited to an ASIC, arguing that would lead the jury to think the specification and its examples were irrelevant to the infringement analysis. Juniper, on the other hand, indeed wanted an instruction limiting the use of the specification and drawings by the jury in forming its infringement conclusions. If you've read my previous reports, you recognize this as the continuing fight between the parties on the issue of how the description and examples given in the patents are to be used in determining infringement.
Palo Alto also objected to a proposed instruction telling the jury that Palo Alto having its own patents is irrelevant to determining whether their products infringe Juniper's patents. The judge was not persuaded by Palo Alto's concerns about that instruction, as she felt it was important to make sure the jury was expressly told to ignore the fact that Palo Alto had its own patents.
Michael Mitzenmacher, Palo Alto's Technical Expert
After the jury was brought in to the courtroom, Palo Alto continued direct examination of its expert, Michael Mitzenmacher, Professor of Computer Science at Harvard, by Michael Jacobs. On Thursday, Mitzenmacher left off in the middle of discussing the '612 patent, so that's where they picked up. Jacobs asked Mitzenmacher about Palo Alto's Network Administrator's Guide "NAG" and its description of the Block IP functionality. Mitzenmacher confirmed for the jury that Block IP is not a dynamic rule because it is setup by the administrator before the machine is put into operation.
Mitzenmacher further opined that the Block IP table, which gets populated with IP addresses during operation, is not equivalent to a rule because, according the '612 patent specification's use of FTP as an example, the Block IP table performs a different function. Further, the '612 patent discusses the use of an ASIC (application specific integrated circuit) to allow ports to be opened or closed on the fly, which thus constitutes dynamic rule creation, whereas Block IP is intended to block traffic from coming in. These differences, according to Mitzenmacher, mean that Palo Alto's accused products are not equivalent to the invention claimed in the '612 patent.
Juniper's attorney, Morgan Chu, then began his cross examination of Mitzenmacher. He started with the '347 "two step" patent, on which Mitzenmacher disputed that the default rule in Palo Alto's system applied to all incoming packets is a default "deny" rule. Mitzenmacher said there is never a second step because there is never any initially denied packet. Chu did get Mitzenmacher to admit that the claim language does not expressly require there be any "action variable", "action value", or "action parameter", all of which mean the same thing. Chu also had Mitzenmacher admit that his experiment to prove that packets in Palo Alto's system are not assigned an initial action value of deny was unnecessary because review of the source code directly was sufficient. Chu next asked Mitzenmacher about the report prepared by Juniper's expert witness, Avi Rubin, Professor of Computer Science at John's Hopkins University, which Mitzenmacher admitted did not say that an action value was set in the initial stage of packet processing in Palo Alto's system. However, Mitzenmacher followed that Rubin "suggested or implied" an action value was set.
Chu then turned to the '612 "dynamic rule" patent, and had Mitzenmacher admit that in his own expert report, he referred to the addition of an IP address to the Block IP table as an "action", and not a "rule". This relates to the court's claim construction of the term "rule" to mean "action," and thus is a favorable admission for Juniper. With respect to Mitzenmacher's experiment showing that various dynamic rules in Palo Alto's system take 10 times longer to execute than adding entries to tables, Chu had Mitzenmacher admit that how fast or slow a rule is implemented is irrelevant to the claim language, as there is no requirement that a "rule" take longer than something that is not a "rule." Mitzenmacher also agreed that there is a Block IP table in Palo Alto's system that uses a data structure.
Chu then sought to impeach Mitzenmacher's credibility. First, following up on the failure to consent to his qualifications as an expert on Thursday, which I thought was a waste of time, Chu had Mitzenmacher admit that he has had no teaching or experience in the field of security. Indeed, Chu pointed out how only a couple of all of the papers Mitzenmacher has authored or co-authored even relate to security. Chu then had Mitzenmacher tell the jury how much he was being compensated for his testimony, another issue I think is unnecessary and irrelevant. Both side's experts are getting paid handsomely for their testimony. I was pleased Palo Alto did not raise the issue with the jury, and felt awkward when Juniper did. I've never heard of a jury saying they didn't like one expert or the other because of their hourly rate. It's like the issue of how rich the corporate representatives are, irrelevant, unnecessary, and just unsavory.
Chu then moved on to the '723 "multiple engine processing" patent, which is by far the most complicated of the three patents-in-suit. He first asked Mitzenmacher if software alone can be considered an "engine", which is the term found in the patent that the court construed to mean "processor." Mitzenmacher said if software is on hardware, then it can be considered an engine, implying the point I made in my earlier reports about software alone not being capable of doing anything at all. Chu then had Mitzenmacher agree that each core contained within the Cavium chip inside Palo Alto's products (there are 16 cores) could be considered a separate CPU (central processing unit). Thus, while one of Mitzenmacher's non-infringement arguments is that Palo Alto's products only have one processor, because they only have one Cavium chip, Chu was impressing on the jury how each chip nonetheless has multiple cores, and that each of those cores is a "central processing unit", with emphasis on the word "processing." Seeing where Chu was trying to go with this point, Mitzenmacher disputed that a CPU could be considered a "processor".
Regarding the experiment Mitzenmacher performed, Chu had him admit that it only showed that the slowpath and fastpath functions could run on the same core, not that they always did. Indeed, Chu had Mitzenmacher admit he chose to run them on the same core and did not perform an experiment to see how often they would naturally run on different cores. Chu also pointed out how fastpath can never run on the same core as slowpath when slowpath runs on cores 2 or 15 for the 4000 series products, because fastpath can only run on cores 3-14 on those machines. This debate relates to whether the jury believes Palo Alto's system must always use different cores, or only sometimes.
On the issue of the doctrine of equivalents for the '723 patent, Chu had Mitzenmacher admit that while the normal definition of "routing" would not include pointers, the court's claim construction did. Chu also had Mitzenmacher admit that one of the documents he had relied on in forming his opinions, called "Day In the Life of a Packet", had omissions, which Mitzenmacher said was expected because it was merely meant as an overview document. Chu then had Mitzenmacher admit that to "associate a tag" does not require a physical connection, like in the real world, where associating a price tag with an item for sale in a store does require a physical connection.
With that, court broke for the weekend, with Mitzenmacher still under cross examination to be continued on Monday, or Tuesday if the court is closed for bad weather on Monday. This provides Chu and his team a substantial advantage, as they get to take the weekend to review all the transcripts from the week and plan what other issues they want to raise with Mitzenmacher still on the stand. It also means that Mitzenmacher is not allowed to speak with his attorneys throughout the weekend until his cross examination is completed.
Like on Monday, Tuesday, Wednesday and Thursday, I began Friday short PANW, as I still remained confident the market misinterpreted the judge's claim construction and summary judgment decision and, as a consequence, is underestimating the risk that Palo Alto faces from the suit. Nothing happened Friday to change my opinion. In fact, because I felt Juniper's cross examination of Mitzenmacher was going so well, I was more confident in my short position. I do not plan to attend the trial next week although I may continue to provide updates and thoughts through my Seeking Alpha column and my Twitter feed.
Additional disclosure: My position as of submission of this article to the Seeking Alpha editors is disclosed. However, legal matters like the one discussed herein are highly dynamic as there can be, for example, new developments in the mentioned matter itself or in the law that applies to the matter at any moment without notice. Thus, the opinions I express herein are my opinions as of submission of this article to the Seeking Alpha editors and I may change my opinions or my position in any related security at any time for any reason. This article was originally available for my clients and to the public for a fee. It is now being made available here for free for the public.