[00:00:00] Speaker 03: like to make a motion and therefore for these purposes Judge Stoll will be presiding. [00:00:04] Speaker 03: I move the admission of Andrea Grace Clark Mills who's a member of the bar and is in good standing with the highest courts of Virginia. [00:00:15] Speaker 03: I have knowledge of her credentials that I'm satisfied that she possesses the necessary qualifications. [00:00:21] Speaker 03: So Gracie's been my law clerk for close to a year now [00:00:25] Speaker 03: And she's been absolutely wonderful. [00:00:28] Speaker 03: I have no reservations. [00:00:30] Speaker 03: She will be a wonderful and productive member of our bar. [00:00:34] Speaker 03: She's lovely to have in chamber. [00:00:37] Speaker 03: She's smart. [00:00:37] Speaker 03: She's thoughtful. [00:00:38] Speaker 03: She's hardworking. [00:00:39] Speaker 03: She's going to be an excellent addition to the bar. [00:00:42] Speaker 03: And really, I feel like I'm swearing in two of you at the moment, which is a first for me. [00:00:51] Speaker 03: I'm happy to be able to make this motion, and I ask my colleagues to take it under advisement and consider whether they'll grant it. [00:00:59] Speaker 02: Judge Shaw, should we grant the motion? [00:01:01] Speaker 02: Yes, supporting papers are all in order. [00:01:04] Speaker 02: We grant the motion. [00:01:05] Speaker 03: Okay, so please raise your right hand and face Jim. [00:01:28] Speaker 03: All right, now let's move on with our first case for today. [00:01:31] Speaker 03: Our first case is a consolidated oral argument. [00:01:36] Speaker 03: It's case 2017-1999 and 2018-1433. [00:01:43] Speaker 03: I understand that the clerk's office has explained to you that we're going to do 20 minutes for each side, but here are both of those cases together since almost all the issues overlap. [00:01:55] Speaker 03: Okay? [00:01:57] Speaker 03: Alright, so Mr. Yang, please proceed. [00:02:01] Speaker 00: Good morning, Your Honors. [00:02:02] Speaker 00: May it please the Court? [00:02:03] Speaker 00: Minkme Yang on behalf of Appellant AC Technologies. [00:02:06] Speaker 00: There are two issues common to these appeals where the Board erred. [00:02:10] Speaker 00: First, it looked at a single word, pieces, in a vacuum without considering the entire claim limitation, redundant storage of pieces of a complete file, in view of the remaining claim language in the specification. [00:02:21] Speaker 00: And second, it found the claims anticipated, even though there's no substantial evidence to show that Revenivich's algorithm actually operates independent of an access. [00:02:29] Speaker 00: I want to start with the second issue and then work my way back to the first. [00:02:33] Speaker 00: Specifically, the error with independent of an access is the board's anticipation determination is based on speculation that Rabinovich's algorithm could operate independent of an access instead of any specific disclosure within Rabinovich that the algorithm actually did operate independent of an access. [00:02:50] Speaker 03: I think that your most difficult point, it seems to me, is Dr. Ratner's testimony, isn't it? [00:02:56] Speaker 03: And whether or not that provides substantial evidence for the board's determination? [00:03:00] Speaker 00: No, Your Honor, because Dr. Ratner's testimony doesn't alter the meaning of the sentence. [00:03:06] Speaker 00: What I mean is the board found in this and several IPRs that independent of an access counts the number of accesses and only copies when the access count is non-zero. [00:03:16] Speaker 00: So in order to be independent of an access, Rabinovich has to disclose excluding host accesses from the access count. [00:03:23] Speaker 00: If Rabinovich discloses excluding host accesses, then it could anticipate. [00:03:27] Speaker 00: But if it doesn't, then it cannot. [00:03:28] Speaker 00: Dr. Ratner's testimony [00:03:30] Speaker 00: specifically is that a specific variable CNT parentheses e comma x sub s is an indicator of future demand and based on this testimony the board concluded that a specific sentence within Rabinovich doesn't just mean that host accesses are counted. [00:03:47] Speaker 00: But the problem with that is just because the variable is an indicator of future demand, that doesn't mean that the sentence doesn't also tell you to count host accesses. [00:03:56] Speaker 00: They're not mutually exclusive. [00:03:57] Speaker 00: The sentence could mean the variable is an indicator of future demand and still tell you to count host accesses. [00:04:04] Speaker 00: And if that's true, then the board couldn't have dismissed this explicit statement to count host accesses. [00:04:09] Speaker 00: But more importantly, no matter how you interpret the sentence, whether you interpret it to mean what it says on its face to count host accesses, or if you interpret it to mean what the board says it means based on Dr. Ratner's testimony, that particular sentence, it doesn't tell you to... [00:04:24] Speaker 00: not count host accesses, either way. [00:04:26] Speaker 02: And that's the critical- Doesn't our case law say that it's not that a reference has to expressly say, hey, we're not going to perform this claim step. [00:04:35] Speaker 02: It's enough that the reference teaches a method in which that step isn't performed. [00:04:44] Speaker 00: Well, Your Honor, the specific claim requires that the steps be independent of an access. [00:04:48] Speaker 00: Right. [00:04:49] Speaker 02: And if you look at the parameters that they're using, and if they're independent of the access of the computer unit, as defined, you know, mapped onto the reference, then isn't that enough? [00:05:01] Speaker 00: Well, it would be if the board hadn't specifically found that the prior art isn't independent of an access, because it counts the number of accesses. [00:05:08] Speaker 02: Access of the computer unit, where the computer unit is the host, right? [00:05:12] Speaker 00: The benefit doesn't explicitly disclose one way or the other whether the host accesses are counted. [00:05:19] Speaker 02: Do you agree though that the law is that you don't have to express, it's okay if the reference doesn't expressly say that it's not performing something so long as it in fact is not performing it? [00:05:36] Speaker 02: My issue is that the reference doesn't... I understand what you're... Your issue is that you don't think there's any support for the idea that there's access, that it's independent of access of the computer unit. [00:05:47] Speaker 02: That's correct. [00:05:48] Speaker 02: Okay, I understand your position. [00:05:49] Speaker 02: I just thought that maybe you were arguing something different with respect to the law. [00:05:53] Speaker 02: Maybe requiring that a negative limitation has to be expressly taught. [00:05:57] Speaker 00: No, that's not what I'm saying. [00:05:59] Speaker 00: That's not what I'm saying at all. [00:06:00] Speaker 00: I'm saying that the negative limitation has to be... [00:06:05] Speaker 00: possible for the negative limitation to be performed. [00:06:07] Speaker 00: In this case, there is no evidence either way whether this limitation is performed or not. [00:06:13] Speaker 00: There has to be some evidence that the limitation is actually performed. [00:06:20] Speaker 00: So if I could use an analogy, let's say I have a recipe for sugar-free brownies, and the prior art discloses a brownie recipe where one of the ingredients is a white powder. [00:06:28] Speaker 00: If the white powder is sugar, then clearly the recipe isn't sugar-free. [00:06:32] Speaker 00: But at the same time, let's say the board said the white powder is baking soda. [00:06:36] Speaker 00: If the white powder is baking soda, that still doesn't disclose sugar-free brownies. [00:06:39] Speaker 00: Sugar or sugar can be one of the various other ingredients. [00:06:45] Speaker 00: And that's the exact same thing we have here. [00:06:49] Speaker 00: Just because the claims require independent of an access, and the board found that Rabinovich doesn't operate independent of access because accounts host accesses. [00:06:58] Speaker 03: But your example actually goes right to the heart of what Judge Stoll was asking you about and what you said you weren't arguing. [00:07:05] Speaker 03: You said if somebody says the white substance is baking soda, that doesn't mean one of the other ingredients couldn't be sugar. [00:07:12] Speaker 03: But if you're talking about a recipe, personally I love to cook. [00:07:16] Speaker 03: I've never seen a recipe that doesn't list all the ingredients that are meant to go into it. [00:07:21] Speaker 03: So if sugar isn't listed as one of the ingredients, but there is in fact a white powder listed and then that white powder is determined to be baking soda, I don't see how that example of yours differs at all [00:07:33] Speaker 03: from the concern that Judge Stoll had, which is that you're requiring the teaching of sugar-free, even in the absence of sugar being listed as an ingredient. [00:07:43] Speaker 00: But in that case, you have to look at the other ingredients. [00:07:45] Speaker 00: And one of the other ingredients has to either disclose sugar or not. [00:07:49] Speaker 00: In this case, you have to look at Rabinovich's algorithm and see whether there's an actual disclosure of not counting hoax taxes. [00:07:58] Speaker 00: And there isn't. [00:07:59] Speaker 02: So where do you think that is in the algorithm? [00:08:02] Speaker 00: I'm saying that there is no disclosure anywhere within the document telling you whether the account host accesses or not. [00:08:15] Speaker 03: Do you want to move on to your pieces, Arjun? [00:08:18] Speaker 00: Before I do that, I'd like to make one clarification about the briefing. [00:08:24] Speaker 00: Appendix 2289 is a diagram that illustrates the copying limitations. [00:08:35] Speaker 00: It's the last page of the appendix. [00:08:37] Speaker 00: On this page, what I want to point out is... Which appendix is this? [00:08:42] Speaker 00: Oh, I'm sorry, Your Honor. [00:08:43] Speaker 00: It's the 1999 appendix. [00:08:47] Speaker 00: Oh, okay. [00:08:47] Speaker 00: And what page again? [00:08:49] Speaker 00: 2289. [00:08:57] Speaker 00: What I want to point out is that the claim requires two separate copying limitations. [00:09:01] Speaker 00: First the computer unit A makes a copy to computer unit B and then a computer unit B makes a copy to computer unit C. Those are marked as C U D S and D S. [00:09:13] Speaker 00: And I also want to point out that Rabinovich itself discloses only a single copying step. [00:09:18] Speaker 00: A single host replicates to another host. [00:09:21] Speaker 00: The final written decision doesn't address this requirement of a second step, except to say that Rabinovich's algorithm can be performed. [00:09:29] Speaker 00: Likewise, on page 44 of the red brief, appellees argue, and I quote, each host in Rabinovich periodically runs replication algorithm. [00:09:37] Speaker 00: But what that means is when a single host runs the replication algorithm, that same host is making copies to other hosts. [00:09:45] Speaker 00: In other words, when host A runs the replication algorithm, host A copies to host B, and then host A continues to run the algorithm and copies to host C, and then host A copies to host D and host A copies to host E and so forth. [00:09:56] Speaker 00: But what the claims require is that host A copy to B and then B copy to C. Clearly, A to B and then A to C and then A to D using Rabinovich's algorithm, that's not the same as A to B and then B to C as claimed. [00:10:08] Speaker 00: And that's another reason why there's no anticipation. [00:10:13] Speaker 00: So if there's no questions, I'd like to move on to the pieces argument. [00:10:22] Speaker 00: The issue with piece is that the board didn't consider the entire claim limitation and certainly didn't consider it in view of the remaining claim language in the specification. [00:10:31] Speaker 00: There's no dispute over the construction of the word pieces itself or what the word means, but what the board's error is of the word pieces is the only thing that it did actually construe. [00:10:40] Speaker 00: It didn't explicitly construe the entire claim limitation, redundant storage of pieces of a complete file. [00:10:46] Speaker 00: And to the extent that it implicitly construed it, that construction is unreasonable because it fails to distinguish redundantly storing pieces of a file, just the pieces, and redundantly storing the entire file. [00:10:57] Speaker 02: Why was the board's reliance on the claim language at least one piece of data stored by a computer unit in a redundant manner? [00:11:06] Speaker 02: Why was that wrong? [00:11:07] Speaker 02: I mean, it seems to me that the board was looking at it and saying, well, it says you have to store at least one piece in a redundant manner. [00:11:17] Speaker 02: So that covers also storing all the pieces in a redundant manner. [00:11:23] Speaker 00: Right, but you still have to store the pieces in a redundant manner. [00:11:28] Speaker 02: It is at least one piece in a redundant manner. [00:11:32] Speaker 00: That's correct, but you have to store the pieces. [00:11:34] Speaker 02: I understand you're later, but okay, your later limit is to say, refer back [00:11:44] Speaker 02: with antecedent basis to the language at least one piece of data. [00:11:49] Speaker 02: So in every time you refer to a piece of data later in the claim, it refers back to at least one piece. [00:11:58] Speaker 02: That's correct. [00:11:58] Speaker 02: So why was the board's construction incorrect under the broadest reasonable construction? [00:12:04] Speaker 00: Because the board's construction failed to distinguish the difference between redundantly storing a piece and redundantly storing the complete file. [00:12:11] Speaker 00: If I could give like an analogy. [00:12:12] Speaker 02: What claim language do you rely on for showing that that's actually recited, where it says at least one piece of data? [00:12:21] Speaker 00: The claims require redundant storage of at least one piece of data. [00:12:24] Speaker 00: So the act of redundant storage must be performed on the at least one piece. [00:12:29] Speaker 02: Where does it say it has to be each piece individually redundantly stored? [00:12:34] Speaker 00: It doesn't say each piece must be redundantly stored. [00:12:36] Speaker 00: It says that a piece must be redundantly stored. [00:12:40] Speaker 02: It says at least one piece. [00:12:41] Speaker 00: Right. [00:12:42] Speaker 00: So the act of redundant storage must be performed on the pieces. [00:12:46] Speaker 03: See, I don't understand that. [00:12:47] Speaker 03: So it says at least one piece. [00:12:50] Speaker 03: It is clearly contemplating multiple pieces. [00:12:53] Speaker 03: So why can't three pieces be redundantly stored together? [00:12:57] Speaker 00: There's nothing that prohibits that in the claim. [00:13:01] Speaker 00: But I'm saying that there's a distinction between redundantly storing pieces and redundantly storing the complete file. [00:13:06] Speaker 00: If I could give an analogy... How? [00:13:08] Speaker 03: Because the one to three that I just told you about... Okay, um... Council, you're being very disruptive. [00:13:23] Speaker 03: Sit back down. [00:13:23] Speaker 03: If you want to take a break and use part of your time to talk to him, that's fine, but I was in the middle of asking him a question, and you thought obviously what you had to tell him was more important than the question I had to ask. [00:13:31] Speaker 00: I apologize, Your Honor. [00:13:33] Speaker 00: If you look at the claims, what the claims say is [00:13:40] Speaker 00: So I'm looking at Appendix 62. [00:13:41] Speaker 00: The claims say that the complete file or the... I've got the patent. [00:13:53] Speaker 03: Is there a particular claim you want me to look at? [00:13:56] Speaker 00: Yes, please look at claim 1 of the 125 patent. [00:13:59] Speaker 03: Okay, I've got that in front of me. [00:14:03] Speaker 00: It says that the computer unit stores one complete file of data. [00:14:06] Speaker 00: That's referring to storing the complete file. [00:14:08] Speaker 00: But then it says each file has a plurality of individual pieces, the pieces containing parts of the file. [00:14:14] Speaker 00: That's referring to the fact that each complete file that's been stored is broken down into a plurality of individual pieces. [00:14:21] Speaker 00: And then the remaining claim language, the first element, is starting with a set of instructions. [00:14:27] Speaker 00: That says at least one piece of data stored by a computer in a redundant manner. [00:14:31] Speaker 00: What that's referring to is that the one piece of data is being stored in a redundant manner, not the complete file. [00:14:36] Speaker 02: Or doesn't it say at least one piece? [00:14:38] Speaker 00: It does, but what has to be redundantly stored is the one piece, not the complete file. [00:14:44] Speaker 03: No, what has to be redundantly stored is the at least one piece. [00:14:46] Speaker 00: That's correct. [00:14:47] Speaker 03: So it could mean two pieces, or the 10 pieces, or all of the pieces could be redundantly stored. [00:14:52] Speaker 00: But you have to redundantly store the pieces, not the complete file. [00:14:57] Speaker 00: Could I please give an analogy? [00:14:59] Speaker 00: Let's say if I told you I was buying a new car. [00:15:01] Speaker 00: In that example, you would expect that I would be driving out of a car dealership with a new car. [00:15:06] Speaker 00: If I showed you a muffler or an engine on a cinder block, that would make no sense to you. [00:15:11] Speaker 03: Likewise, if I told you that I was buying parts of a new car, then in that example, you would understand that... And if I put a pie on the table in front of you, a whole pie, and I said, feel free to take at least one piece with you, and you took the whole pie with you, [00:15:27] Speaker 03: You would have taken at least one piece with you. [00:15:29] Speaker 00: Yes, but I would have taken them in pieces. [00:15:31] Speaker 00: I wouldn't have taken the whole pie. [00:15:33] Speaker 03: Well, that wasn't my hypo. [00:15:34] Speaker 03: You took the whole pie in my case. [00:15:36] Speaker 03: You were a little piggy. [00:15:36] Speaker 03: What can I tell you? [00:15:38] Speaker 00: I did, but I took the pie in a piece-by-piece fashion. [00:15:41] Speaker 00: I didn't take the pie in a... [00:15:45] Speaker 02: Can I ask you something else moving on to the specification? [00:15:48] Speaker 02: I was a little puzzled by a block quote in your brief. [00:15:52] Speaker 02: It's on page 26. [00:15:54] Speaker 02: And here I think that you were trying to make the point that [00:15:58] Speaker 02: You know, the specification really is just talking about pieces. [00:16:02] Speaker 02: And it's the block quote you have on page 26. [00:16:06] Speaker 02: But if I look at the full quote, you have some dot, dot, dots in here. [00:16:11] Speaker 02: And the full quote actually says that the invention is based on distribution of the entire data GD or the fields F to several cells. [00:16:23] Speaker 02: But you deleted the part where it says entire data GD. [00:16:27] Speaker 02: I mean, I'm confused by this because I think this disclosure in column eight might suggest that you have redundant storage either of the fields or of the entire data GD. [00:16:40] Speaker 00: Right, but if you look at figure two, figure two explains how the embodiments work. [00:16:47] Speaker 00: You're taking the entire data quantity GD and dividing it into little pieces. [00:16:52] Speaker 00: And then you're storing those little pieces in the cells. [00:16:56] Speaker 00: It doesn't matter that you may store all the little pieces. [00:16:59] Speaker 00: What we're doing is you're storing the little pieces on a piece-by-piece basis. [00:17:02] Speaker 02: The problem I'm having is your claims are written broader than that. [00:17:04] Speaker 02: They say storing at least one piece redundantly. [00:17:07] Speaker 02: And then the specification includes some support for the idea that you wanted to have coverage, not just of redundantly storing fields F, but as it's in quoted column 8, lines 13, the entire data GD. [00:17:24] Speaker 02: So I understand there might be one embodiment where there's just fields, but you also seem to have wanted coverage [00:17:30] Speaker 02: of the storing the entire file or entire data GD redundant. [00:17:34] Speaker 02: I don't think that's true, Your Honor, because if you look at Column 8, tell me how I'm misreading this language. [00:17:42] Speaker 00: I think what column 8 is referring to is that the result of this process is redundantly storing the entire file. [00:17:49] Speaker 00: But the way that the process works is it claims a specific implementation in which you divide the complete file into pieces and then you store those pieces. [00:17:57] Speaker 00: You may store the complete file at the end of the process, but the way the process works is that you're storing them on a piece-by-piece basis. [00:18:06] Speaker 03: Okay, you're into your rebuttal time. [00:18:07] Speaker 03: Do you want to save the remainder? [00:18:08] Speaker 00: I would like to save the remaining time for rebuttal. [00:18:12] Speaker 03: Thank you. [00:18:25] Speaker 01: Good morning, Your Honor. [00:18:26] Speaker 01: May it please the Court, Dan Chavodian of Perkins Cooey for Appellees Amazon and Blizzard, and with me today are Mr. Wing Liang and Mr. Dan Bagatelle. [00:18:38] Speaker 01: And I'll start with the pieces limitation since that's where the argument left off with Mr. Yang. [00:18:47] Speaker 01: Your honors are correct that the claim says you store or you transfer at least one piece. [00:18:56] Speaker 01: So it can be more than one piece that is transferred and stored. [00:19:00] Speaker 01: And in fact, the preamble says that the computer unit stores the complete file and it has the pieces of the file. [00:19:09] Speaker 01: So that language in the preamble also supports that [00:19:13] Speaker 01: The use of the word pieces does not preclude a device having the complete file as well. [00:19:20] Speaker 03: And we're operating still at this point under our broadest reasonable construction standard, correct? [00:19:25] Speaker 03: That's correct. [00:19:26] Speaker 03: So the broadest, it certainly is a broader construction to say at least one piece can include the entire thing than it would be to say it has to be piece by piece, right? [00:19:34] Speaker 01: That's correct, Your Honor. [00:19:35] Speaker 03: So if the PTO's construction, even if I don't agree with it in a de novo sort of world, [00:19:42] Speaker 03: where I'm trying to find the Phillips construction under the broadest reasonable construction. [00:19:47] Speaker 03: So long as it's reasonable, a law supported by the spec, I have to leave it alone. [00:19:51] Speaker 01: That's correct, Your Honor. [00:19:53] Speaker 01: And the board found that there was nothing in the specification that limits how the transfer occurs, how the storage occurs or any of the mechanics that [00:20:06] Speaker 01: that AC has argued would somehow limit this to just storing just one piece as opposed to storing multiple pieces or storing pieces of parts of a file. [00:20:17] Speaker 01: And Dr. Ratner testified that when the pieces are stored... Well, you say that there's nothing, but Figure 2 actually does that, right? [00:20:24] Speaker 03: Figure 2 does articulate embodiments where the pieces are stored. [00:20:33] Speaker 03: and then they go F to Z in a redundant generator by cell by cell. [00:20:38] Speaker 03: So you said nowhere in the spec does it disclose that, but doesn't it actually disclose that in column eight? [00:20:42] Speaker 01: Well, it doesn't preclude. [00:20:44] Speaker 01: I'm saying nothing precludes storing the pieces as parts of storing the whole file. [00:20:48] Speaker 01: So for example, in figure two. [00:20:50] Speaker 02: Is there anything that expresses the desire of doing that? [00:20:53] Speaker 01: Of doing what you're on? [00:20:55] Speaker 02: Of storing the entire file, redundantly. [00:20:57] Speaker 01: Well, I believe the portion of the specification that Your Honor pointed to at column 8 says that you can redundantly store either the file or the pieces, where it refers to the entire data, GD. [00:21:10] Speaker 03: Well, boy, you might want to walk that statement back, because if you can redundantly store the file or the pieces, and those are two totally different things, then doesn't claim one, which says at least one piece, not include the whole file. [00:21:23] Speaker 03: The way you're using those words, you just made them seem like very distinct and different things. [00:21:27] Speaker 01: Well, I may have misspoken, Your Honor, then. [00:21:29] Speaker 01: As Dr. Ratner explained, when you store the pieces of a file, when the transfer occurs using the TCP IP protocol, the file is transferred in pieces. [00:21:43] Speaker 01: And those pieces are then stored on the receiving device. [00:21:47] Speaker 01: And they're organized, as Dr. Ratner stated, as bytes on the receiving device. [00:21:51] Speaker 01: And bytes are pieces of the file. [00:21:53] Speaker 01: So you do have pieces stored. [00:21:56] Speaker 01: Even when you store the complete file, it's stored in pieces organized as bytes. [00:22:01] Speaker 01: And so the specification actually doesn't use the word file anywhere. [00:22:05] Speaker 01: It doesn't use piece of a file anywhere. [00:22:08] Speaker 01: So the closest thing we have to that is the entire data set GD and the fields F. And so the fields F, I believe, are the closest thing in there to a byte. [00:22:21] Speaker 03: I'm going to ask you to leave the courtroom if you do it again. [00:22:23] Speaker 03: I can't listen to him because you're talking so loudly that I'm unable to follow his argument, and that's really unfair. [00:22:29] Speaker 03: If you do it again, I will ask you to leave the courtroom. [00:22:35] Speaker 03: Please continue. [00:22:37] Speaker 01: Certainly. [00:22:42] Speaker 01: I'm sorry I just derailed you, but I apologize for that. [00:22:47] Speaker 03: I couldn't follow what you were saying, so I'm sorry. [00:22:49] Speaker 01: In fact, I just wanted to point out that we also have dependent claims 3 and 10 that [00:22:56] Speaker 01: recite sending and receiving and storing pieces until the complete file has been transferred. [00:23:02] Speaker 01: So that shows that the first claim can't preclude that. [00:23:05] Speaker 01: There's nothing in the first claim that can be construed in a manner that would somehow prevent the transferring and storing of all of the pieces constituting a complete file. [00:23:21] Speaker 03: Do you want to address, please, the independently anticipation argument? [00:23:26] Speaker 01: Certainly, Your Honor. [00:23:41] Speaker 01: So I can sort of set the context for this. [00:23:45] Speaker 01: There is an original host in Rabinovich that stores a file. [00:23:51] Speaker 01: It then transfers it in pieces over to a second host. [00:23:54] Speaker 01: So if we can consider the first one the original host, I'll call the second one host 2. [00:24:00] Speaker 01: And then Rabinovich says that every host runs this replication algorithm. [00:24:05] Speaker 01: So it's not just the original host that would be copying to other hosts, but Revenovitz says every host autonomously runs the replication algorithm. [00:24:14] Speaker 01: So host 2 would then run this algorithm, and if host 2, when the factors are right, such as host 2 has too much of a load, it would then offload the file to another host as well, or replicate the file to another host. [00:24:29] Speaker 01: So that it does in fact disclose the second replication that's required by the claims because every host in Rabinovich that receives a copy of the file would then run the replication algorithm and copy to another host when the factors are correct. [00:24:44] Speaker 03: And that would be independent of the original transferring computer. [00:24:47] Speaker 01: That's correct, Your Honor. [00:24:48] Speaker 03: And that's where the disclosure comes, you think? [00:24:51] Speaker 01: And that's the primary issue here, is whether that second replication from host two to host three would be independent of an access of the original host. [00:25:01] Speaker 01: And one thing that AC argues broadly in the brief and here today about host accesses, and what is precluded is not a host access, it's an access of the original host. [00:25:16] Speaker 01: And the factors that are disclosed in the algorithm that are evaluated are accesses of what are referred to today as host two or accesses directed to host two that pass through host three. [00:25:30] Speaker 01: And that's the variable that's discussed in Rabinovich and in our briefs as count, it's abbreviated CNT and then parentheses E comma X sub S. [00:25:46] Speaker 01: And what that variable means is it's a count of the number of requests directed to host 2 that pass through host 3. [00:25:55] Speaker 01: So host 2 is overloaded with work. [00:25:58] Speaker 01: It needs to replicate a copy of this file so another server can take on part of the burden of servicing clients. [00:26:06] Speaker 01: And it looks to where would be a good location to place this new replication of the file. [00:26:13] Speaker 01: And if most of the requests are coming to host 2 through host 3, host 2 knows that would be a good location. [00:26:20] Speaker 01: I'll put a copy at host 3, and that can then take on a lot of the burden. [00:26:24] Speaker 01: And so that's what that variable represents. [00:26:27] Speaker 01: And it's explained in both Rabinovich and Dr. Ratner also discussed this at Appendix 644. [00:26:35] Speaker 01: And he explained that it's a predictor of future demand for copies of the file. [00:26:41] Speaker 01: So that's how Host 2 knows that Host 3 would be a good location. [00:26:45] Speaker 01: And none of these variables depend on an access of the original host. [00:26:51] Speaker 01: As Your Honor's noted, the algorithm sets forth what we can call the recipe. [00:26:56] Speaker 01: It lists everything that is used to decide whether to replicate the file. [00:27:02] Speaker 02: So if something's not listed... Five parameters, am I right in understanding there's kind of five parameters that it's looking at? [00:27:08] Speaker 01: There's a number of parameters. [00:27:11] Speaker 01: The one argued [00:27:13] Speaker 01: In AC's brief is that count E X sub s that I referred to there's another one that's count S X Which is the number of accesses of the second host are those the only access parameters? [00:27:26] Speaker 01: I believe that's correct your honor and in deciding that [00:27:34] Speaker 01: Then Dr. Ratner explained how that's a prediction of the demand on the host 3, the possible future demand on host 3. [00:27:43] Speaker 01: And so that's why host 3 would be a good location. [00:27:47] Speaker 01: And all of the parameters are set forth in the algorithm. [00:27:52] Speaker 01: And what isn't disclosed there is not a factor that's relied upon. [00:27:58] Speaker 01: As Your Honor's noted, [00:28:00] Speaker 01: you say what the algorithm relies upon, then things not listed are not relied upon, are excluded. [00:28:07] Speaker 01: And counsel said today they're not arguing that [00:28:12] Speaker 01: that the algorithm needs to say what it excludes, but that's actually exactly what they argued in their reply brief. [00:28:19] Speaker 01: They underlined X in exclude, and they say Rabinovich doesn't say it excludes that factor, the access of the original host, when in fact it's not listed anywhere in the algorithm, so it is excluded. [00:28:35] Speaker 03: Is there anything else that you feel like you need to cover? [00:28:38] Speaker 01: I don't believe so, Your Honor. [00:28:46] Speaker 00: Thank you, Your Honor. [00:28:47] Speaker 00: I just want to first briefly address the anticipation issue. [00:28:50] Speaker 00: The claims require a receipt of a copy and then making a copy based on the received copy. [00:28:56] Speaker 00: But a benefic doesn't disclose receiving a copy and then replicating. [00:28:59] Speaker 00: All it does is disclose having a copy and then replicating. [00:29:02] Speaker 00: The claims, however, require receiving that copy first, which is the part that's not disclosed by a benefic. [00:29:11] Speaker 00: And then going back to the pieces issue, the first thing I want to point out is that the preamble distinguishes a piece from a complete file. [00:29:20] Speaker 00: These are two separate things. [00:29:24] Speaker 00: If I could direct you to claim one of the 125 patents, it explicitly says that [00:29:34] Speaker 00: It stores one complete file, and separately from that, that one complete file is divided into pieces, and then those pieces are redundantly stored. [00:29:41] Speaker 03: I'm sorry, what column and line number are you referring to? [00:29:44] Speaker 00: I'm sorry, column 27, starting at line 30. [00:29:54] Speaker 00: It says that the computer unit stores at least one complete file of data. [00:29:57] Speaker 00: That's the complete file that is restored. [00:30:00] Speaker 00: And then the complete file is divided into individual pieces. [00:30:04] Speaker 00: And then the remaining claim steps store that piece of the complete file, not the complete file itself. [00:30:13] Speaker 00: And I also want to point out that the distinction that Mr. Shavodian is making undermines the entire purpose of this invention. [00:30:19] Speaker 00: If you were to say that you were redundantly storing the complete file and it redundantly stores the pieces just because each file has bits and bytes, then [00:30:29] Speaker 00: The normal course of redundantly storing complete files would meet all of these claim limitations. [00:30:33] Speaker 00: In that case, there would be no reason to draft the claims in such a way that it specifically says that the claims operate on a single piece and not on the complete file. [00:30:44] Speaker 00: It seems to me that when you're redundantly storing a single piece, what you're doing is you're storing that particular piece. [00:30:50] Speaker 00: You may be storing other pieces as well, but the operation of the redundant storage occurs on the piece, not on the complete file. [00:30:57] Speaker 00: What you're doing is you're dividing the file into plurality pieces and considering the operation of the step of redundant storage on each individual piece. [00:31:08] Speaker 00: And that's exactly what is illustrated by claim two. [00:31:11] Speaker 00: I'm sorry, figure two. [00:31:14] Speaker 00: In figure 2, you have the entire file gd, but instead of storing gd itself 1 through 10 in each of the cells, it divides gd into pieces 1 through 10, and then it looks at each individual piece. [00:31:28] Speaker 00: It looks at each individual piece, and then it puts, for example, piece 8 into these specific pieces. [00:31:33] Speaker 00: into each of these specific cells. [00:31:35] Speaker 00: Now, plausibly, it could look at other pieces as well and put them into these cells, but what it's doing is it's looking at the pieces. [00:31:42] Speaker 00: It's not looking at the complete file. [00:31:44] Speaker 00: What Rabinovich does is it doesn't look at the pieces at all. [00:31:47] Speaker 00: It's just a file that's been copied over and over again. [00:31:50] Speaker 00: The file may have bits and bytes, but you're not looking at the bits and bytes when you're making the copies. [00:31:54] Speaker 00: You're looking at the complete file. [00:31:56] Speaker 03: So your view is that this claim requires [00:31:58] Speaker 03: that the file have to be in pieces, so even if the file was just cut in half, such that there were two halves, that would be two pieces, and then if each of those pieces were stored, even though it would result in the complete file being stored, that would satisfy the claim, but Rabinovich doesn't break it up into anything, it just stores the whole thing. [00:32:18] Speaker 03: Is that basically what your argument is? [00:32:20] Speaker 00: My argument is more than that. [00:32:21] Speaker 00: My argument is that after you break it up, you have to look at the pieces. [00:32:27] Speaker 00: Revenevich not only doesn't break up any files, it doesn't look at any particular piece. [00:32:32] Speaker 03: What part of the claim requires you to look at the pieces? [00:32:35] Speaker 03: I thought we were talking about at least one piece of data stored by a computer unit in a redundant matter. [00:32:41] Speaker 00: Right. [00:32:42] Speaker 00: The redundant manner, when you're storing a, I'm sorry, I'm almost out of time. [00:32:46] Speaker 00: Can I finish the sentence? [00:32:47] Speaker 00: Yes, please. [00:32:48] Speaker 00: When you're storing a piece in a redundant manner, what you're doing is you're looking at the piece and storing that piece in a redundant manner. [00:32:54] Speaker 00: You're not looking at the complete file and making duplicates of the complete file. [00:32:57] Speaker 00: You're looking at the specific piece and making sure that specific piece is being redundantly stored. [00:33:03] Speaker 03: Okay. [00:33:03] Speaker 03: Thank you. [00:33:04] Speaker 03: Thank you, Your Honor.