`
` UNITED STATES PATENT AND TRADEMARK OFFICE
`____________________________________
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________________________________
` DELL INC., EMC CORPORATION, HEWLETT-PACKARD
` ENTERPRISE CO., and HP ENTERPRISE SERVICES, LLC,
`Petitioners,
`v.
`REALTIME DATA LLC,
`Patent Owner.
`____________________________________
`Case: IPR2017-00179
`Patent No. 9,054,728
`Case: IPR2017-00176
`Patent No. 7,161,506
`____________________________________
`
`CROSS-EXAMINATION OF:
`DR. CHARLES D. CREUSERE
`Friday, August 4, 2017
`
`Reported by:
`SUSAN L. CIMINELLI
`Job no: 19344
`
`TransPerfect Legal Solutions
`212-400-8845 - Depo@TransPerfect.com
`
`Realtime Ex. 2005
`NetApp v. Realtime
`IPR2017-01660
`
`
`
`1
`C O N T E N T S
`2 DR. CHARLES D. CREUSERE
`3
`EXAMINATION BY:
`4
` Counsel for Patent Owner
`5
` Counsel for Petitioners
`
`Page 4
`
`PAGE
`5
`146
`
`INDEX TO EXHIBITS
`*There were no exhibits marked at this deposition.
`
`67
`
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`Page 2
`
` DR. CHARLES D. CREUSERE, called for
`cross-examination by counsel for Patent Owner,
`pursuant to notice, at the offices of Winston &
`Strawn, LLP, 1700 K Street, N.W., Washington, D.C.,
`before SUSAN L. CIMINELLI, CRR, RPR, a Notary Public
`in and for the District of Columbia, beginning at
`9:38 a.m., when were present on behalf of the
`respective parties:
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`Page 3
`
`Page 5
`
`1
`P R O C E E D I N G S
`2 Whereupon,
`3
` DR. CHARLES D. CREUSERE,
`4
`was called as a witness by counsel for Patent Owner,
`5
`and having been duly sworn, was examined and
`6
`testified as follows:
`7
`CROSS-EXAMINATION
`8
` MR. SOMMER: On behalf of Petitioner, you
`9
`have Andrew Sommer from Winston & Strawn. With me on
`10
`the phone today is Tom Brown from Dell EMC.
`11
` MR. NOROOZI: And for Patent Owner, Kayvan
`12
`Noroozi.
`13
`BY MR. NOROOZI:
`14
`Q. Dr. Creusere, good morning. I see you
`15
`have some documents in front of you, is that right?
`16
`A. Correct.
`17
`Q. And could you just go through and tell me
`18
`what you have there?
`19
`A. Sure. I have the Franaszek patent,
`20
`Exhibit 1004. Sebastian patent, I can't read the
`21
`exhibit number on this one. The Aakre patent. The
`22
`Hsu paper from Software Practice and Experience. The
`2 (Pages 2 to 5)
`
`A P P E A R A N C E S
`On behalf of Patent Owner:
` KAYVAN B. NOROOZI, ESQUIRE
` Noroozi, P.C.
` 1299 Ocean Avenue
` Suite 450
` Santa Monica, California 90401
` 370.975.7074
` kayvan@noroozipc.com
`
`On behalf of Petitioners:
` ANDREW R. SOMMER, ESQUIRE
` Winston & Strawn, LLP
` 1700 K Street, N.W.
` Washington, D.C. 20006-3817
` 202.282-5000
` asommer@winston.com
`
`ALSO PRESENT:
`
` Tom Brown, Esquire, In-house Counsel EMC
` (Via telephone)
`* * * * *
`
`1
`2
`3
`4
`5
`6
`
`78
`
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`TransPerfect Legal Solutions
`212-400-8845 - Depo@TransPerfect.com
`
`
`
`Page 6
`1
`Fallon patent '506. The Fallon patent '728. A copy
`2
`of my declaration for the '506 patent. And a copy of
`3 my declaration for the '728 patent.
`4
` Q. Okay, thank you. How did you prepare for
`5
`today?
`6
` A. I reviewed all of the materials that I had
`7
`used in preparing my declaration. I reviewed my
`8
`declaration. I reviewed the decision to institute by
`9
`the Patent Board. I reviewed the Patent Owner
`10
`response. And I had discussions with Drew and
`11 Michael Woods about this material.
`12
` Q. Did you speak to anybody else?
`13
` A. No.
`14
` Q. Did you consider or review any other
`15 materials besides the one that you just mentioned?
`16
` A. I did review my transcripts from the --
`17
`one of the depositions I did back in January. And I
`18
`do not -- I do not recall reviewing other
`19
`documentation. But I could have missed something.
`20
` Q. Other than what you just told me, do you
`21
`recall reviewing any prior art documents,
`22
`dictionaries or other documents that you had not
`Page 7
`
`1
`previously cited in your declaration?
`2
` A. I believe that in studying the Sebastian
`3
`prior art, I did take a look at very brief look at
`4
`patent -- patents cited by Sebastian. Well, it's
`5
`given on column 4, line -- it's cited on column 4,
`6
`line 18 in the Sebastian patent and it's cited as an
`7
`application by Mr. Schindler, 08/970,220. I did take
`8
`a very brief look at that.
`9
` Q. What caused you to want to look at that
`10
`reference?
`11
` A. Based on discussion, I was curious exactly
`12 what that reference entailed.
`13
` Q. And why?
`14
` A. Because Sebastian refers to it when
`15
`discussing possible compression that might, possible
`16
`specific compression algorithms that might be
`17
`applied, and so I felt I had not previously looked at
`18
`it. I felt that it might be worth looking at.
`19
` Q. Okay. So what were the column and line
`20
`numbers again?
`21
` A. It is column 4, line 18.
`22
` Q. And on your original declaration, you did
`
`TransPerfect Legal Solutions
`212-400-8845 - Depo@TransPerfect.com
`
`Page 8
`not consider or discuss the Schindler reference cited
`in the Sebastian reference, true?
` A. No, that's true.
` Q. And did your evaluation of the Schindler
`reference in any way modify or influence your
`opinions compared to the opinions set forth in your
`declaration?
` A. No, it has no impact on my opinions.
` Q. About how long did you prepare for this
`deposition?
` A. I believe -- well, we spent yesterday in
`discussions for most of the day. And I put in
`probably about six or seven additional hours prior to
`the discussion.
` Q. Let me ask you about Franaszek. In
`Franaszek, regardless of whether the system
`recognizes the data type, has data type information,
`representative samples of each block are tested to
`select an optimal encoder for the block, right?
` A. So Franaszek first does a comparison to
`see if type information is available. If it is, it
`uses that type information to select a list of
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`Page 9
`1
`possible encoders that is optimized for that type.
`2
`If not, it will select from the default encoder list.
`3 And in both cases, once it's finished with that
`4
`process, it will test all of the decoder, all of the
`5
`encoders on that list. It will test a -- all the
`6
`encoders on that list on a sample of the block of
`7
`data and it will choose one of those encoders based
`8
`on that test.
`9
` Q. And as a part of the testing that happens
`10
`in Franaszek, regardless of whether there is a data
`11
`type provided to Franaszek's system or not, the
`12
`testing will always identify the compressibility of
`13
`the data block using the different encoders that are
`14
`tested on the sample, right?
`15
` A. I wouldn't phrase it exactly that way. I
`16
`would say that Franaszek will determine the encoder
`17
`that achieves the highest, that -- let me rephrase
`18
`that. Franaszek will attempt to determine the
`19
`encoder that achieves, that will achieve the highest
`20
`compression on that block. It will not necessarily
`21
`succeed, but it will attempt to do that.
`22
` Q. The purpose of Franaszek's -- withdrawn.
`3 (Pages 6 to 9)
`
`
`
`Page 10
`1
` One purpose of Franaszek's testing on a
`2
`data block sample is to determine the compressibility
`3
`of the data block using the different possible
`4
`encoders that are in the list, right?
`5
` A. Again, I don't believe I would phrase it
`6
`that way. I would phrase it as Franaszek is trying
`7
`to determine which encoder will hopefully compress
`8
`the data block best.
`9
` Q. In order to do that, doesn't Franaszek
`10
`identify for each encoder in the list that is being
`11
`considered, how well that encoder is expected to
`12
`compress the data block?
`13
` A. I would again prefer to phrase it in the
`14
`way that I phrased it, which is that -- that the real
`15
`goal is to determine which decoder, which encoder
`16
`will compress the data block or to try to determine,
`17
`try to estimate which encoder will compress the data
`18
`block the best. I believe that is the goal.
`19
` Q. How does Franaszek's testing and sampling
`20
`approach make that determination?
`21
` A. Franaszek's testing and sampling approach
`22 makes that determination by testing a portion of the
`Page 11
`block with each of the compression algorithms in that
`list, and assuming that at least one of those
`algorithms achieves sufficiently high compression,
`high enough compression to clear a threshold. Then
`the best -- then the -- one of those encoders from
`that list which achieves the best compression on that
`sample will be selected and used to encode the entire
`block.
` Q. And when you say best, you mean highest
`compression ratio, right?
` A. In the preferred embodiment of Franaszek,
`it will choose the encoder that achieves the highest
`compression ratio.
` Q. And is there any other embodiment in
`Franaszek that provides an alternative to what you
`just described?
` A. There is no other embodiment that I have
`seen that I can recall that is explicitly spelled out
`in Franaszek that would do something different than
`that.
` Q. So in Franaszek, the ultimate compression
`technique that's applied to a data block will always
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`TransPerfect Legal Solutions
`212-400-8845 - Depo@TransPerfect.com
`
`Page 12
`be selected based on how well that compression
`technique is expected to compress the data block in
`terms of compression ratio. True?
` A. Again, in the preferred embodiment of
`Franaszek, as spelled out in the description of the
`invention, Franaszek will always choose solely based
`on which of the encoders in its list achieves highest
`compression ratio, again, assuming that encoder
`achieves above the 30 percent threshold.
` Q. When Franaszek doesn't have a data type,
`it will use a default list of compression techniques,
`true?
` A. Yes. Franaszek will use a default list if
`it does not have a data type.
` Q. But Franaszek will not ever select a
`particular compression technique to apply to a data
`block simply because the data, a block does not come
`with a data type, true?
` MR. SOMMER: Object to form.
` THE WITNESS: So you're asking -- could
`you rephrase that question, please?
`BY MR. NOROOZI:
`
`Page 13
`
` Q. Happily.
` A. Okay.
` Q. In all instances, when a data block that's
`provided to Franaszek's system does not come with
`data type information, Franaszek will select the
`ultimate compression technique if one is selected at
`all from a list of possible default compression
`techniques. True?
` A. Yes. Franaszek will do its testing
`procedure on the sample. And based on testing
`procedure, it will -- and assuming that it clears the
`threshold requirement, it will select one of those
`encoders from that list.
` Q. And so when a data block comes into
`Franaszek's system with that data type information,
`it is not possible to predict without any other facts
`which specific compression technique will be used to
`compress that data block, assuming some compression
`technique will be selected. True?
` MR. SOMMER: Object to form.
` THE WITNESS: I wouldn't necessarily say
`that it is not possible to predict, because there are
`4 (Pages 10 to 13)
`
`
`
`Page 14
`1 means that people develop for trying to predict
`2
`compressibility of different blocks without doing a
`3
`full compression. But within the framework of the
`4
`Franaszek patent and his preferred embodiment, his
`5
`embodiment does not -- does not detail, to my
`6
`recollection, a means for doing that, for predicting
`7
`the compressibility prior to doing the sampling.
`8
`BY MR. NOROOZI:
`9
` Q. Now, when Franaszek has data type
`10
`information, it will generate a preselected list of
`11
`compression techniques to choose from for that data
`12
`block, right?
`13
` MR. SOMMER: Object to form.
`14
` THE WITNESS: I would rephrase that and I
`15 would say that if Franaszek has type information, it
`16 will use, it will use the appropriate list of
`17
`compression algorithms for that type. It will not
`18
`necessarily generate it, and there is nothing in
`19
`Franaszek that says it generates the list on the fly.
`20
`BY MR. NOROOZI:
`21
` Q. Okay. So your point is simply that
`22
`Franaszek will have in some fashion a list of
`Page 15
`1
`algorithms that have been predetermined to be the
`2
`appropriate set for a particular data block given the
`3
`data type information that has been provided for that
`4
`data block?
`5
` A. Yes. Franaszek will have a list of
`6
`algorithms associated with a given type of data,
`7
`assuming that, assuming that it knows that type of
`8
`data. Franaszek doesn't address the situation
`9
`directly, where, where it doesn't know -- where there
`10 might be type description data. It doesn't know it.
`11
`But assuming it knows that data, it will have some
`12
`sort of a list associated with that data of possible
`13
`compression algorithms and it will select, it will go
`14
`through the same process of selecting one of those
`15
`compression algorithms to apply to encode the block.
`16
` Q. And to your point just a second ago,
`17
`Franaszek does not contemplate a situation in which
`18
`it is provided with data type information but does
`19
`not have a list of compression techniques associated
`20
`with that data type. True?
`21
` A. From the preferred embodiment of
`22
`Franaszek, I don't recall any indication that
`
`Page 16
`
`1
`Franaszek contemplates that scenario.
`2
` Q. When Franaszek has a data type information
`3
`and has generated a preselected list of encoders for
`4
`that data type or has identified that list, let's say
`5
`-- let me withdraw and start over because I know you
`6
`don't like the phrasing "generate." Withdrawn.
`7
` When Franaszek has a data type and has
`8
`identified a preselected list of encoders for that
`9
`data type, it will sample and test all those
`10
`algorithms against the data block like we said
`11
`earlier. True?
`12
` A. Yes, it will do the -- it will take a
`13
`sample that will compress that sample each of the
`14
`encoder types. It will do the threshold test and it
`15
`will select the encoder type that chooses the highest
`16
`compression that exceeds the threshold.
`17
` Q. So in the '728 patent specification, it's
`18
`taught that a data block will be compressed with
`19 multiple different encoders and the ultimate
`20
`compressed block that's output will be the one that
`21
`has the highest compression ratio. True?
`22
` A. In the '728 patent, I certainly would
`
`Page 17
`agree that there is at least one embodiment within
`the '728 patent that operates in such a manner where
`it, where it, where it attempts to choose, select an
`encoder that achieves the highest compression rate.
`I'm not certain that that is the only -- I'm not
`certain that the '728 patent doesn't have additional
`embodiments or options that might not do things a
`little bit differently.
` Q. The '728 patent does not teach a testing
`and sampling approach, whereby a sample of the data
`block is tested in order to ultimately select the
`compression technique that's used for the data block.
`True?
` A. Well, I believe it depends on how you
`define sample. I mean, the '728 patent certainly
`discusses compressing the entire data block. And one
`could contend that the entire data block is just a
`sample of 100 percent.
` Q. You don't have an opinion like that in
`your declaration, right?
` A. No.
` Q. And --
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`TransPerfect Legal Solutions
`212-400-8845 - Depo@TransPerfect.com
`
`5 (Pages 14 to 17)
`
`
`
`Page 20
`1
` A. I always thought it was Hsu, but I was
`2
`never good with Chinese names so I'm not absolutely
`3
`sure.
`4
` Q. I'll call it Hsu. I don't know either.
`5 Withdrawn.
`6
` You agree that the Hsu reference describes
`7
`a precompression phase and a compression phase?
`8
` A. Yes. The Hsu reference describes a
`9
`precompression phase followed by the compression
`10
`phase. Yes.
`11
` Q. And during the precompression phrase, the
`12
`data type and the compressibility of each data block
`13
`is determined. True?
`14
` A. As laid out in Hsu's paper, the
`15
`precompression, the precompression phase has
`16
`basically four parts you could say. One part is
`17
`using the new file routine to determine, to come up
`18
`with a determination of the data type. And then the
`19
`other three parts are to calculate the three
`20
`redundancy metrics -- three redundancy metrics, and
`21
`those four pieces are used to select the compression
`22
`algorithm.
`
`Page 18
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
` A. At least not to my knowledge.
` Q. And you recognize that in the embodiment
`in the '728 patent that we are discussing, where the
`entirety of a data block is compressed using multiple
`different compression techniques, the compressed data
`block that is ultimately output will be selected
`after compression has already occurred, right?
` A. Let me just phrase that. So if what
`you're asking -- so my interpretation of how the '728
`patent operates, I believe, and I believe the way a
`person of ordinary skill in the art would interpret
`it is that '728 patent in certain modes of operation
`at least will encode the entire block with the
`possible encoders, and will then select the
`appropriate encoded entire block based on a
`compression, a test of the compression test. Test on
`which one achieves the highest compression. So if
`that's what you're asking me, then that's -- you
`know, that's my interpretation of how that will work.
` Q. Well I think I'm asking you something
`slightly different, which is, in the '728 patent's
`teachings, the compression technique that's used for
`Page 19
`1
`the data block that's output, the compressed data
`2
`block is not selected by the system before
`3
`compression is complete. True?
`4
` MR. SOMMER: Object to form.
`5
` THE WITNESS: The block that's going to be
`6
`output is selected by the system on the basis of how
`7 will it was compressed. The entire block. How will
`8
`the entire block compress. In Franaszek, a sample
`9
`block is tested and then that block is selected and
`10
`then that block is fully compressed. So that's my
`11
`interpretation of the relative differences between
`12
`how Franaszek operates and how the '728 patent
`13
`operates.
`14
`BY MR. NOROOZI:
`15
` Q. And when you were talking about Franaszek,
`16
`you said that block will be selected. I think you
`17 meant that compression technique will be selected?
`18
` A. Sorry. I meant that compression technique
`19
`will be selected on the basis of how that compression
`20
`technique compresses that sample of the block.
`21
` Q. You agree that the Chu reference --
`22
`reference, do you pronounce Chu -- oh, Hsu?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`Page 21
` Q. Hsu's approach will always determine some
`information about the compressibility of each data
`block, right?
` MR. SOMMER: Object to form.
` THE WITNESS: Yes. I believe that that,
`those four pieces of information, the type and the
`three compressibility metrics as he called them or
`redundancy metrics he calls them. I would have to
`refer to the paper. Those certainly provide
`information about the block and how well, and how it
`should be compressions.
`BY MR. NOROOZI:
` Q. And in Hsu's approach, of those four
`pieces of information that the data type and the
`three redundancy metrics, it will always be the case
`that at least one, and likely more of those will be
`identified for a given data block, right?
` A. As regards the three redundancy metrics,
`those are basically statistical analysis
`calculations, and those can always be calculated for
`any data block. As far as the -- as the data type
`information as output by Hsu's new file routine,
`6 (Pages 18 to 21)
`
`TransPerfect Legal Solutions
`212-400-8845 - Depo@TransPerfect.com
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`Page 22
`that, that's -- that will -- that should in my, in my
`opinion, as a person of ordinary skill in the art, my
`best estimate is that will always output something.
`It will always give you something. Now exactly what
`it outputs under certain circumstances, I'm not
`totally sure of.
` Q. Sebastian teaches a system that uses
`different filters for different data types, right?
` A. Sebastian uses the terminology of filter
`which is a non-standard terminology in the field to
`represent his terminology -- this is his terminology
`for basically data compression algorithms. So
`Sebastian has a set of filters that are designed for
`specific data types, which if it recognizes that data
`type it will apply, and then it has a generic filter
`that will apply in situations where it doesn't,
`either does not have or does not recognize an
`incoming data type.
` Q. Now, to the last part of your statement,
`in your declaration, you don't identify any teaching
`from Sebastian that says that it will apply the
`generic filter to circumstances where it does not
`Page 23
`1
`have or does not recognize the data type. True?
`2
` A. You're asking me about what I'm saying in
`3 my deposition.
`4
` Q. Your declaration?
`5
` A. I'm sorry, I apologize, my declaration?
`6
` Q. Yes.
`7
` A. I would have to review that. I have a
`8
`great deal of discussion of Sebastian. If you want
`9
`to talk about a specific statement I make, that would
`10
`be perfectly fine but there is a great deal of
`11
`discussion of Sebastian in the motivation to combine
`12
`sections of the both of the two declarations, so I
`13 would have to review that. I'm not -- I'm not sure
`14
`off the top of my head.
`15
` Q. All right. You agree that in Sebastian,
`16
`the teaching is that the system identifies or has a
`17
`data type, and then attempts to apply a particular
`18
`compression algorithm that has been associated with
`19
`that data type, as long as that compression algorithm
`20 which it calls a filter is installed in the system?
`21
` A. I think I would need to really take a look
`22
`at Sebastian here because Sebastian is very -- it
`
`TransPerfect Legal Solutions
`212-400-8845 - Depo@TransPerfect.com
`
`Page 24
`1
`only mentions this a couple of times. It's very
`2
`particular about his language on this.
`3
` MR. SOMMER: I think it's at the bottom of
`4
`your pile. It was the first one you grabbed. Yes.
`5
` THE WITNESS: I'd like to be as accurate
`6
`as possible in this.
`7
`BY MR. NOROOZI:
`8
` Q. Sure. Let me try to direct you to a place
`9
`in Sebastian unless there is a place that you're
`10
`already going to?
`11
` A. Well, there are two places in Sebastian
`12 where this is discussed, if I recall. But if you
`13 would like to direct me to the particular one you're
`14
`interested in, that's fine.
`15
` Q. So go ahead. Tell me what you identify in
`16
`Sebastian?
`17
` A. So first place where it's mentioned is
`18
`column 1, starting at line 55. Which states that if
`19
`a filter is installed which matches the format of the
`20
`data to be encoded the advantages of format specific
`21
`compression can be realized for the data. Otherwise
`22
`a generic filter is used that achieves performance
`Page 25
`similar to other non-specific data compression
`systems. So in this statement, and then there is one
`other statement further on, it discusses whether or
`not a filter is installed that matches the format of
`the data to be encoded.
` Q. And let me just ask you right there. You
`agree that that at least suggests that there is
`already a data type that has been identified or
`provided. True?
` A. It would appear from this, from this part
`of the statement, to suggest that. Though I don't
`recall any discussion from this patent on the process
`of identifying data types, and whether or not data
`types might not have an identifier. I don't recall
`that being discussed in this patent. It could be in
`here somewhere but certainly the statement refers to
`a test of whether or not there is a filter installed
`that matches the form last of the data to be encoded,
`so it does, it does seem -- it does imply that one
`knows what the format of the data to be encoded is.
` Q. And so I think the next place you were
`going to look is column 4, around line 9. Is that
`7 (Pages 22 to 25)
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`Page 26
`
`1
`right?
`2
` A. Yes. Correct. That's the other place
`3 where this is mentioned.
`4
` Q. And you see that in column 4, lines 9
`5
`through 23, Sebastian teaches that the filter
`6
`selection system 22 receives source data, and checks
`7
`to see if any of the filters installed in its system
`8
`or the compression techniques installed in its system
`9
`support the source data format.
`10
` A. Yes, they do.
`11
` Q. And source data format you understand to
`12 mean data type?
`13
` A. I believe within this framework, source
`14
`data format would be equivalent to data type.
`15
` Q. And so again, you understand that suggests
`16
`-- withdrawn.
`17
` And so again, you understand that
`18
`Sebastian is at least suggesting there that it
`19
`already has the data type information and is simply
`20
`looking for the appropriate compression algorithm in
`21
`its system. True?
`22
` A. I would -- it's not clear to me. I mean,
`Page 27
`1
`Sebastian doesn't really discuss that. It says the
`2
`filter selection system receives source data format
`3
`and checks selection criteria, the very selection
`4
`criteria there. All that is -- all that is is the
`5
`selection criteria is just a test to see if it
`6 matches the formats of one of these filters that's
`7
`installed on the system.
`8
` That's certainly -- because the process
`9
`of -- of determining -- that certainly implies that
`10
`system knows what the data source format is, but
`11
`again, there is no discussion of how the system would
`12
`know that other than if, well there is no discussion
`13
`at all, period, of how the system would know that.
`14
`Other than in some very perhaps simple way, like with
`15
`a dot suffix or something like that. And there is no
`16
`discussion of what would happen if the system didn't
`17
`know if it couldn't make a decision because there was
`18
`no such information available.
`19
` Q. So Sebastian does not have any discussion
`20
`of a scenario in which the data type for a data block
`21
`is not available to it and what it would do in that
`22
`scenario. True?
`
`TransPerfect Legal Solutions
`212-400-8845 - Depo@TransPerfect.com
`
`Page 28
`1
` A. Not to my recollection. Sebastian is a
`2
`long patent. To be able to say that definitively, I
`3 would have to go through line by line. But to my
`4
`recollection, that is not something that is -- that
`5
`is covered in Sebastian.
`6
` Q. And column 4 in that same section you were
`7
`looking at, and throughout Sebastian, the only
`8
`scenario in which Sebastian teaches using its generic
`9
`filter is when its system does not already have the
`10
`desired or appropriate compression technique for a
`11
`particular data type. True?
`12
` A. Sebastian specifically states that if it
`13
`doesn't have a filter that matches the source data's
`14
`format, then it will use a generic filter. So that
`15
`is as far as it goes on the subject. It doesn't, to
`16 my recollection, state other scenarios.
`17
` Q. With respect to the Aakre reference, you
`18
`rely on that reference for the proposition that Aakre
`19
`teaches realtime compression, right?
`20
` A. That is correct.
`21
` Q. And specifically what Aakre teaches is
`22
`using controllers and buffers to make data
`
`Page 29
`1
`transmission to a storage device into a continuous
`2
`stream so that the storage device isn't starting and
`3
`stopping during the storage of a data block, right?
`4
` A. Aakre specifically focuses on the -- on
`5
`the application of a tape storage drive. That's
`6
`correct. And Aakre's focus is to -- is to create a
`7
`continuous stream of data so that the data can stream
`8
`out to the device in realtime, and it does -- as
`9 Aakre points out, that has the benefit of keeping the
`10
`data, keeping the tape from starting and stopping so
`11
`the data can be read continuously on to the tape, as
`12
`it flows out of the system.
`13
` Q. You mean written continuously?
`14
` A. I'm sorry, written continuously on the
`15
`tape as it flows out of the system.
`16
` Q. In your declaration, you don't identify
`17
`any teaching from Aakre about compressing the data
`18
`itself within a certain amount of time. True?
`19
` A. The concept of compressing the data in a
`20
`certain amount of time is imply its in Aakre because
`21
`for it to achieve the realtime criteria of being able
`22
`to pass data from its input through its compressor
`8 (Pages 26 to 29)
`
`
`
`Page 30
`1
`through its buffer through the system on to the tape
`2
`at such speed as the tape is capable of writing. So
`3
`it needs to pass the data through fast enough so that
`4
`the tape can write the data on to the tape as it's
`5 moving without stopping, without halting. That does
`6
`imply that there has to be a time requirement in how
`7
`fast that job can be completed. Otherwise it would
`8
`fall behind and the tape would have to stop.
`9
` Q. Before I get into the substance of what
`10
`you just stated, that opinion that you just gave me
`11
`is not in your declaration, right?
`12
` A. The opinion I believe is in my
`13
`declaration, but the -- but the exact level of detail
`14
`I've given you just now is not, to my recollection,
`15
`spelled out in exactly that manner.
`16
` Q. In your declaration as a starting point,
`17
`you don't provide any opinion on the appropriate
`18
`construction of realtime or realtime compression,
`19
`true?
`20
` MR. SOMMER: Object to form.
`21
` THE WITNESS: It's my recollection that --
`22
`are you referring to claims construction?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`Page 31
`
`Page 32
`
`Right?
` A. And a compressor. Compressor, controller
`and a buffer.
` Q. All right. And a buffer is a means for
`holding on to a certain amount of information. So
`that you don't have to pass it immediately through to
`the storage device. Right?
` A. I would say that a little bit differently
`because one could say that a memory is a means for
`holding on to a piece of data. A buffer is a memory
`unit that is designed as temporary space. Typically,
`it's called first in first out, so data comes in one
`side, goes out the other side, in the order in which
`it came in.
` And a buffer is fundamentally designed to
`be temporary. Buffers generally have the ability to
`accommodate, to allow the amount of information
`inside the buffer to fluctuate with time so the
`buffer is used to smooth the flow of data coming out
`of the compression algorithm so that it can be flow
`evenly at -- uniformly into the tape drive unit.
` Q. And what about the controller? What does
`Page 33
`
`1
`BY MR. NOROOZI:
`2
` Q. Yes.
`3
` A. No, I do not recall providing an opinion
`4
`on claims construction on the terminology of
`5
`realtime.
`6
` Q. Okay. And you don't have an opinion in
`7
`your declaration that Aakre teaches compressing the
`8
`data itself within a particular period of time.
`9
`True?
`10
` MR. SOMMER: Object to form. Go ahead.
`11
` THE WITNESS: Again, as I said briefly
`12
`earlier, I don't discuss specifically Aakre
`13
`compressing within some period of time, but I discuss
`14
`Aakre's intended application, which is the
`15
`application of streaming continuously to a tape drive
`16
`so that it does not have to start and stop, and that
`17
`enforces the realtime constraint based on how fast
`18
`the tape drive can write information on to its
`19 magnetic tape.
`20
`BY MR. NOROOZI:
`21
` Q. Well, let me explore that for a second
`22
`because Aakre teaches using controllers and buffers.
`
`1 Aakre teach about the controller?
`2
` A. Aakre teaches that the controller, if you
`3
`look at Aakre's patent, the most germane figure in
`4
`the patent is figure 1 of the patent. And most
`5
`fundamentally, Aakre teaches that the controller has
`6
`control lines going both to the compressor and to the
`7
`buffer. It's very critical that the controller
`8
`control the compressor, adjusting its output relative
`9
`to the buffer full