`IPR2014-00949, Paper No.27
`November 3, 2015
`
`trials@uspto.gov
`
`571-272-7822
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`VMWARE, INC., INTERNATIONAL BUSINESS MACHINES
`CORPORATION and ORACLE AMERICA, INC.,
`Petitioners,
`
`v.
`
`ELECTRONICS AND TELECOMMUNICATIONS
`RESEARCH INSTITUTE,
`Patent Owner.
`____________
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`____________
`
`Held: August 28, 2015
`____________
`
`
`
`BEFORE: BRIAN J. McNAMARA, MIRIAM L. QUINN, and
`GREGG I. ANDERSON, Administrative Patent Judges.
`
`
`
`
`
`The above-entitled matter came on for hearing on Friday, August
`28, 2015, commencing at 1:31 p.m., at the U.S. Patent and
`Trademark Office, 600 Dulany Street, Alexandria, Virginia.
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`
`
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`KATHERINE KELLY LUTTON, ESQUIRE
`
`
`MICHAEL RUECKHEIM, ESQUIRE
`
`
`LEERON G. KALAY, ESQUIRE
`
`
`Fish & Richardson, P.C.
`
`
`500 Arguello Street
`
`
`Suite 500
`
`
`Redwood City, California 94063
`
`
`
`ON BEHALF OF PATENT OWNER:
`
`
`MATTHEW C. PHILLIPS, ESQUIRE
`
`
`DEREK MEEKER, ESQUIRE
`
`
`Renaissance IP Law Group LLP
`
`
`9600 SW Oak Street
`
`
`Suite 560
`
`
`Portland, Oregon 97223
`
`
`
`
`
`
`
`
`
` 2
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`
`
`
`P R O C E E D I N G S
`- - - - -
`JUDGE McNAMARA: Good afternoon, everyone.
`This is -- this case is VMware, Inc., International Business
`Machines, Corporation and Oracle America, Inc., Petitioners,
`versus Electronics and Telecommunications Research Institute,
`Patent Owner, Case IPR2014-00901, which has been joined with
`IPR2014-00949.
`I'm Judge McNamara. Judge Quinn and Judge
`Anderson will be participating remotely.
`Beginning with the Petitioner, could I have the parties
`please introduce themselves.
`MS. LUTTON: Thank you, Your Honor. Kathy Lutton
`from Fish & Richardson for VMware. Joining me, are Leeron
`Kalay and Michael Rueckheim from Fish & Richardson and our
`client from VMware, Brooks Beard.
`JUDGE McNAMARA: Thank you. Patent Owner?
`MS. LUTTON: We also have co-counsel here.
`JUDGE McNAMARA: I'm sorry.
`MR. FRIEDMAN: Todd Friedman from Kirkland &
`Ellis and I'm here on behalf of Petitioners IBM and Oracle and
`my client, Steven Purdy and Peter O'Rourke, in the back of the
`room and Ben Lasky from Kirkland & Ellis is also here.
`
` 3
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`
`JUDGE McNAMARA: Thank you. Anyone else?
`MS. LUTTON: No, Your Honor.
`JUDGE McNAMARA: Okay. And from Patent
`Owners?
`MR. PHILLIPS: Good afternoon, Your Honor.
`Matthew Phillips here with Derek Meeker from the Renaissance
`IP Law Group firm on behalf of the Patent Owner.
`JUDGE McNAMARA: Okay. Thank you. Welcome
`to the Patent Trial and Appeal Board.
`Each party will have 45 minutes of total argument time.
`The Petitioner will proceed first, present its case in chief and then
`after that the Patent Owner will respond to the Petitioners' case
`and, finally, the Petitioner can use any time it has reserved for a
`rebuttal.
`There's no motions or anything else that are going to be
`addressed at this hearing, so is everybody ready to proceed?
`MS. LUTTON: We are, Your Honor.
`JUDGE McNAMARA: All right. Well, let's begin
`with the Petitioner.
`Is there some amount of time you'd like me to alert you
`
`to?
`
`MS. LUTTON: Yes, Your Honor. I would like you to
`reserve 15 minutes, please.
`JUDGE McNAMARA: Okay.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 4
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`
`MS. LUTTON: In addition to that, since we do have
`multiple Petitioners, I would respectfully request a break before
`our rebuttal so that we can confer and I can make sure I'm
`communicating with views of the group, just a short five-minute
`break.
`
`JUDGE McNAMARA: Let me consider it and see
`where we are at that time.
`MS. LUTTON: Okay. Thank you, Your Honor.
`Your Honor, we have a hard copy of the slides. Can we
`hand it up?
`JUDGE McNAMARA: Yes. I'd be happy to take it.
`MS. LUTTON: Okay. Great.
`JUDGE McNAMARA: One thing I should remind
`everybody about during the presentations today -- you can just
`pass that up -- please make sure you refer to the slide number that
`you're talking about so that the judges who are participating
`remotely will be able to look at your -- at the correct slide that
`you want them to be looking at.
`MS. LUTTON: We will, Your Honor. And just as a
`preliminary matter, the version of this slide deck we handed up is
`in a little bit different order than the version we filed.
`JUDGE McNAMARA: Have there been any changes?
`MS. LUTTON: There have been no changes
`substantively to the slides. We have used three of Patent Owner's
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 5
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`slides that were very high level slides just to illustrate a few
`points, but all the slides are ones that were filed.
`JUDGE McNAMARA: Okay. Let's go.
`MS. LUTTON: Also as a preliminary matter, this Court
`has already considered the Hathorn reference with respect to
`anticipation and has determined that the '346 patent is not
`anticipated by Hathorn. In that determination, the Court was
`restrained by the rules of anticipation.
`In this case what the panel is being asked to decide
`today is whether the patent claims of the '346 patent are rendered
`obvious in light of Hathorn in view of Mylex. So it's a different
`test, and Petitioners would submit that all of the claims of the '346
`patent are obvious in view of Mylex and Hathorn.
`Also as a preliminary matter, there were a couple of
`issues raised in the briefing that we believe the Court doesn't
`necessarily need to decide if it determines that the art of Hathorn
`and Mylex is analogous. So for those two issues, which I'll
`describe in a second, the Petitioners plan on relying on the
`briefing, but we are happy to answer any questions the panel has.
`The first issue is the construction of RAID and the
`second issue is whether Hathorn teaches a RAID if the term
`RAID is construed in the same way that the Board construed it in
`a Dell IPR. Both of those issues are only relevant to the issues of
`whether the art is analogous and whether one of skill in the art --
`one of skill in the art would combine the two references and we
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 6
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`believe there's ample evidence that whether or not Hathorn
`discloses a RAID that the references are combinable.
`I'm going to start first with slides 3 through 5.
`The Petitioners have submitted three slides that map
`Claims 1 and Claims 9, the two independent claims of the '346
`patent to Figure 4 of the patent just to give a brief overview of the
`patent before we start into the argument.
`Turning to slide 4, I'm just going to give a brief
`overview of the '346 patent. As the panel may recall, the '346
`patent relates to fault tolerant redundant storage systems. Even
`Dr. Conte, who is Patent Owner's expert, in his definition of
`person of ordinary skill in the art used the term fault tolerance as
`it relates to mass storage devices. So there's no question as to the
`field of art that the patent resides in.
`In terms of what the patent teaches, the Patentee
`disclosed a couple versions of prior art and described its
`invention in terms of the differences between the invention and
`the prior art. I've shown here on the slide, on slide 4, Figure 2
`from the prior art on the left and on the right Figure 4 from the
`patent, which purports to demonstrate the invention.
`As the panel can see, Figure 2 has most all of the same
`elements -- actually all of the same elements as Figure 4. There
`are only some slight differences.
`Figure 2 includes two host computers, which are
`connected to a RAID. In the prior art, that RAID included
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 7
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`redundancy. It included two RAID controllers, each of which
`had two controllers within them. Those RAID controllers were
`connected up through switches and hubs to the host computers.
`Figure 4 has the same components. There are a number
`of host computers that are connected to RAID devices through a
`series of switches and hubs. The only difference between the
`prior art and what Patentee contends is its invention is the
`communication paths between the RAID controllers and the
`switches, and this is illustrated in Figure 4 in the red.
`So essentially what the Patentee contends it added to the
`art was getting rid of the direct path between the RAID
`controllers, which can be seen in Figure 2. There's a direct path
`between the two, the two lines in the middle. The Patentee's
`purported invention is getting rid of that direct path and, instead,
`using the communication channel or the information flow shown
`in red on the right-hand side.
`Turning to slide 5, so the independent Claims 1 and 9
`are identical in the sense that the only thing the Patentee
`purported to add was those red communication channels. Claim 9
`has one additional limitation that Claim 1 does not have and that's
`the concept of fault tolerance, but there's no suggestion that fault
`tolerance in and of itself is new, so this is not something that was
`added to the patent as an invention. It's just an additional feature
`of Claim 9.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 8
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`
`Turning to the next slide, which is Patent Owner slide
`3 -- so there are only three slides I'm going to refer to that the
`Patent Owner submitted. They're very high level, 3, 5 and 4 in
`that order.
`The reason I chose this slide is because the language
`that's highlighted in red in this slide is the language that
`corresponds to the red arrows that I just illustrated demonstrate
`what the Patentee purported to add to the prior art. This language
`is in this slide 3 in Claim 1.
`And if you turn to claim -- or to -- in slide 3 if you turn
`to slide 5, that same language appears in Claim 9.
`And if you turn to Patent Owner's slide 4, you can see
`that the Patent Owner is presenting the same type of picture that I
`previously showed showing the added communications in red.
`So, again, this is all that was added purportedly by the
`patent was getting rid of a direct connection between the RAID
`controllers and, instead, replacing that direct connection with
`connections between the network interface controllers and the
`switches.
`Now, let's turn to the prior art.
`I'd like to start with Mylex, which is on the next slide,
`slide 7. There's no dispute that Mylex discloses all of the same
`elements of the patent, all of the same components, and it has all
`the same paths.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 9
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`
`The only thing that's disputed is the information flow
`and what I've provided here is a mapping of the patent claims, the
`independent Claims 1 and 9 to Mylex. And as this shows, Mylex
`has multiple hosts, the two green hosts that are shown in the
`figure. It has a RAID at the bottom, which is shown in purple,
`and it has two RAID controllers, controller 0 and controller 1,
`each of which have two NICs. Those two NICs are connected to
`the hosts through a series of hubs and switches, which is shown in
`the yellow.
`Turning next --
`JUDGE McNAMARA: Just NICs, you mean N I Cs?
`MS. LUTTON: N I Cs, that's correct.
`JUDGE McNAMARA: To make sure we have the
`record right. Thank you.
`MS. LUTTON: Thank you, Your Honor.
`Turning to slide 9, in addition to having all the same
`elements that are in the patent claim in terms of the physical
`components, the other thing that Mylex discloses is the fault
`tolerance failover that is in Claim 9.
`So the only thing that is missing from Mylex -- the only
`thing that's missing from Mylex is the portions in red that I
`showed you in Patent Owner's slides, which is the communication
`between the various NICs and the switches. That's the only thing
`that purportedly was added by the patent and is the only thing that
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 10
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`the parties are disputing is in the combination of Mylex and
`Hathorn or is taught by the combination of Mylex and Hathorn.
`What is missing is actually disclosed in Hathorn. If we
`turn to the next slide, which is slide 10, the communication
`channels that were not overtly disclosed in Mylex are expressly
`disclosed in Hathorn, and you can see here that not only on the
`right -- and I'll step through this in a little bit more detail -- not
`only are those channels disclosed in Hathorn, Hathorn teaches
`how to modify the prior art, which had direct connections to add
`those type of communication channels.
`And if we go two slides down in the deck, which is
`slide 12. So I know this is a test in eye reading. There's a lot
`here, but I wanted to demonstrate the differences or the -- more
`importantly, the similarities between the patent and Hathorn.
`So in the patent -- and these are the two figures we
`previously discussed. In the patent on the left-hand side, upper
`left, you see the prior art which has a direct communication
`channel between the RAID controllers and you can see what the
`patent purports to add are these communication channels between
`the different NICs.
`In particular, NIC 1 is communicating with NIC 4,
`which is actually the third one over. It's a little out of order. But
`NIC 1 is communicating with NIC 4 and NIC 2 is communicating
`with NIC 3. That's what the patent teaches is new.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 11
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`
`What Hathorn discloses is that you can take the prior
`art, like the prior art that Hathorn discloses itself, like the patent
`prior art, like Mylex, you can take that prior art and you can
`replace the direct channel between the RAID controllers shown
`on the bottom left, which I believe is Figure 2 of Hathorn, and
`you can replace that direct channel with the exact communication
`channel that the patent teaches, which is NIC 1 communicating
`with NIC 4 and NIC 2 communicating with NIC 3.
`JUDGE McNAMARA: And why would someone do
`that without having -- without looking at the patent?
`MS. LUTTON: Somebody would do this because what
`Hathorn is showing is it's showing -- it's teaching that that's what
`you would do. It's teaching the prior art that has a direct channel
`and then it's saying there are reasons you would take that direct
`channel and you would replace it with the communication
`channel on the right.
`For example, one reason might be cost. As we all know
`when you're building infrastructure, you want to have fewer lines,
`fewer devices. Getting rid of a line is going to save money.
`That's one reason.
`Also redundancy. If you look at the channel on the left,
`there's only -- or the picture on the left, there's only one
`connection between the NICs on the left and the NIC on the right.
`If you have a single connection, there's much more likely to be
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 12
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`failure than if you have multiple connections like you have in
`Figure 3.
`So for redundancy and cost purposes, one would be
`inclined to look at Hathorn, realize it teaches that you can change
`the communication paths and can achieve both of those benefits.
`JUDGE McNAMARA: Is that suggested in your
`references?
`MS. LUTTON: That is suggested in the references and
`I believe it's suggested in Hathorn itself. I have this a little
`further on in the slides and I'll just move to it now.
`So turning to slide 36, Dr. Horst, the Petitioners' expert,
`offered an opinion on the interpretation of Hathorn and said one
`of ordinary skill in the art would have found it obvious to modify
`the system disclosed in the Mylex patent so the expense of a
`heartbeat path is avoided by modifying the network interface
`controlling units to exchange information with each other as
`claimed in the '346 patent.
`JUDGE QUINN: But the question from my colleague
`was where in Hathorn do we have that motivation or that
`suggestion. We know that Dr. Horst has said this, but we're
`looking for what he just said, the suggestion in Hathorn.
`MS. LUTTON: So the suggestion appears at column 7
`starting at line 56 and ending at column 8, line 15.
`Do we have a slide on that or should I --
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 13
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`
`JUDGE McNAMARA: Ms. Lutton, column 7 you're
`talking about is Hathorn, right?
`MS. LUTTON: Hathorn, that's correct, and I can read
`the relevant --
`JUDGE QUINN: That section that you're referring to,
`column 7 and 8 referring to the reduction of communication links,
`is that where you're referring to?
`MS. LUTTON: That's correct.
`JUDGE QUINN: Isn't there an operational difference
`here with respect to Hathorn that motivation makes sense with
`respect to Hathorn because of Hathorn being a circuit path rather
`than a logical path or a packet switching type of network, like the
`fibre would be, and that the network in Hathorn had controllers
`that were spaced apart in distance such that you couldn't use the
`typical packet switching that you would have.
`MS. LUTTON: So both Hathorn and Mylex and the
`patent disclose fiber channel networks. They're disclosing fibre
`channel networks in part because fibre channel can be used over
`distance. For example, in Mylex -- and I'll give you the reference
`here. Mylex talks about the fact that fibre channels are used so
`that you can span over distance as taught in Hathorn.
`So if we turn to slide 21. So there's two -- there's a
`couple of references in Mylex about distance. There is a
`reference in Mylex, which is not on this slide, that talks about
`being able to use Mylex over a campus-wide network. Actually
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 14
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`that is in this one. There's two references. The campus-wide
`network reference is in this one.
`Mylex says, fibre channel is an ideal SAN interconnect
`because it provides scaleable performance with virtually
`unlimited addressing and can span campus-wide distances. So
`both Mylex and Hathorn talk about using fibre channels because
`you're going to go over distance.
`JUDGE QUINN: And my question is, with regards to
`your contention that one would modify Mylex to include this
`feature that you're referring to in Hathorn, you're purporting to
`modify the heartbeat path in Mylex with this new path of Hathorn
`that goes through the fibre channel switch.
`What is the compatibility issue with that, wouldn't that
`be -- what operational issues would you have with that, because
`the Patent Owner has raised significant issues that we need to
`consider regarding whether that is a -- that a person of skill in the
`art would be motivated to change out the existing heartbeat path
`with something Hathorn does not have.
`MS. LUTTON: So I think that there are a couple issues
`in the argument that Patent Owner advances. One is that Patent
`Owner seems to advance the theory that you need to physically
`combine the two references. And if you can't physically take the
`ports from one and add them to another that one would not think
`to combine them. That is not the test under the law and I believe
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 15
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`it's In re Keller that makes it very clear that the test for
`combinability is not a physical combinability test.
`So the issue is really would you take the concept that is
`taught in Hathorn and would you apply that to Mylex and that
`concept, as you noted, is the idea of getting rid of a direct path
`and using previously existing paths to transmit a message, in this
`case, as you noted, the heartbeat.
`So right now in Mylex the way it's designed, the
`heartbeat transmits from one RAID controller to the other. There
`are a few issues with that. First of all, if there is a break in that
`one line, nobody will know whether that -- the error occurred in
`one of the two RAID controlling units or whether the error
`occurred in the line because there's no redundancy.
`If you actually apply the Hathorn teachings to Mylex
`and actually remove that line and send the heartbeat over the fibre
`channel, you now have a redundant network. So that's one reason
`why one of skill in the art might apply the concepts taught in
`Hathorn to Mylex. Another --
`JUDGE QUINN: I have a problem with that argument,
`though, because heartbeats are a typical -- typically heartbeats are
`a different signal altogether than a rewrite command from a host.
`So to say that you can easily just change out communication and
`now extend the latency of a heartbeat by sending it through a
`fibre channel switch, we need to see a motivation to do that and
`whether that would, indeed, be a predictable modification.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 16
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`
`MS. LUTTON: So in terms of the heartbeat path, so
`there are a couple of reasons why you would do that. You're right
`that there is a direct path and you would be changing that through
`the other path. In terms of the latency -- one second.
`So there are a number of benefits to changing that that
`would have to be weighed. One, as I mentioned, is redundancy.
`So if you have a failure in that heartbeat path, you would actually
`know what is failing and you don't have to worry about whether it
`was actually the heartbeat line that failed.
`Additionally, if you are sending it through the fibre
`channel, you're actually sending the heartbeat signal through all
`the things that you care about failing, because the data is coming
`through those same channels.
`If you're sending the heartbeat channel directly between
`two units, then the failure could be at any point and you're not
`actually testing the path the data flows in. Also, one would be
`motivated to move it. Because if you move that communication,
`the direct communication up into the host-site network, you can
`transmit more than just the heartbeat path.
`Yes, you're right that part of the thinking is you have the
`path to transmit the heartbeat path, but you could transmit other
`information as well once you have those pathways established.
`JUDGE ANDERSON: Isn't Mylex redundant already?
`It has a failover path. Isn't that the redundancy you're talking
`about?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 17
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`
`MS. LUTTON: So Mylex has all the redundancy that
`the prior art taught and the patent has. It has the multiple RAID
`controllers, the multiple NICs, but the heartbeat path itself is just
`a single path. It's not a redundant path that you would achieve if
`you actually went through the host-site network where you would
`have two different pathways by which that heartbeat message
`could be sent if you're talking about the heartbeat message, or if
`you're sending other information there would be two pathways to
`send it on.
`So there's a failure at any one of those pathways with
`the Hathorn application. The Mylex you have redundancy. If
`you just rely on Mylex alone, you don't have that redundancy.
`JUDGE ANDERSON: You may have answered this
`before, but where is Hathorn -- now let's turn to Hathorn. Where
`is it talking about redundancy?
`MS. LUTTON: So if we turn to the picture of Hathorn,
`which is slide 10, Hathorn is actually disclosing redundant paths
`between the NICs. As you can see, it's not a single line that goes
`from the storage controllers, in between the storage controllers.
`There's actually redundant paths that go through switches and
`back and forth to the different NICs. So one of skill in the art
`looking at this would realize that there are redundant paths.
`JUDGE ANDERSON: Does Hathorn describe this any
`more than what the drawing figure shows as being redundant in
`the case of a failure?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 18
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`
`MS. LUTTON: So Hathorn does talk about
`redundancy. When Hathorn is talking about redundancy, it's
`talking about redundancy in terms of the controllers for backup
`systems. So it does stress the need for redundancy. The whole
`purpose of the Hathorn reference is to have redundant storage that
`space at a point where you can store information to the RAID
`controllers. And if there's an error with one RAID controller,
`then the system continues to operate and you can still access your
`data.
`
`So it talks generally at a high level about redundancy
`and the need for redundancy. I don't believe it specifically
`addresses redundancy with regard to the specific lines
`themselves. It talks about it as an overall concept and then it
`discloses a picture that shows the redundant connections.
`JUDGE QUINN: I have a question for you about
`something you said earlier. You were saying that by putting the
`heartbeats through the fibre switch that you also can detect link
`failures in addition to controller failures. Is that what you said?
`MS. LUTTON: So my point was if you send -- so that
`is what I said. If you send the heartbeat through the different
`NICs, then if you have a failure on a NIC, then you're going to
`detect that with the heartbeat. If you send it directly between the
`two RAID storage controllers, you may or may not be able to
`detect that. So it would actually fail at the point where you're
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 19
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`actually have a failure that's going to matter to your system,
`which is through the NICs and through the storage controllers.
`JUDGE QUINN: But sending that signal through the
`fibre switch, you could also be picking up failure to the ports or
`the cables or the switch itself, not necessarily the heartbeat, so I'm
`not sure why would one take out a heartbeat signal that is so
`simple as a cable between the devices and complicate it to where
`now you don't know if it's truly a heartbeat failure or if this -- that
`the cable is bad or the link between them is bad or the 351 bridge
`between the two switches is bad.
`MS. LUTTON: So, first of all, I want to go back to the
`point that Hathorn is not just talking about changing the
`communication paths for the heartbeat. It's a broader concept that
`you've got a more robust system for a number of reasons if you
`change the communication paths and remove the direct
`communication paths and replace it with the paths that are shown
`in Hathorn. So it's not just about the heartbeat.
`In terms of the heartbeat, no matter which path you send
`it on, you're always going to have the issue that something else is
`failing in that path other than what you're trying to detect. The
`benefit of going through the fibre channel network as opposed to
`having a single communication line between the two -- between
`the two devices, there's a couple of benefits.
`One, you can have multiple paths, so you can detect the
`heartbeat on two paths and not one. So if you have a failure in
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 20
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`one, but the other path is good, then you know that there's
`actually no failure of what you're trying to detect a failure in.
`You don't have that benefit if you're going directly between the
`two.
`
`The other issue is --
`JUDGE QUINN: Do you have any testimony on this
`particular issue or any evidence whatsoever that the heartbeats
`would accomplish what you just said?
`MS. LUTTON: There is testimony from Dr. Horst, and
`this is on slide 44, where Dr. Horst says, one of ordinary skill in
`the art would have been motivated to use the Hathorn
`configuration to achieve the redundancy goals set forth in the
`Mylex paper. So he's speaking to this issue of redundancy.
`JUDGE QUINN: No, I'm talking about do you have
`anything specific about these benefits you just spoke about?
`MS. LUTTON: I don't have it in the level of detail that
`I just spoke about it. All of the papers, including the patent,
`Mylex and Hathorn speak at a high level about redundancy
`concerns and wanting to ensure that there are redundant
`connections, so I have it at that level. I don't have it at the level
`of detail that I was -- I was giving examples. I don't have
`testimony at that level of detail.
`JUDGE QUINN: Okay. Because my problem here is
`that you're providing argument concerning what one of skill in
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 21
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`the art would have believed to be beneficial, but we don't have
`any testimony supporting that.
`MS. LUTTON: We don't. We have testimony
`supporting the fact that one of skill in the art would apply the
`teachings of Hathorn to Mylex because of redundancy and cost
`reasons, which is the level of detail that was provided in the
`papers and the level of detail also that is provided in the patent.
`You're right, it --
`JUDGE QUINN: Okay. So can you identify for me,
`then, what are the specific motivations that you do have evidence
`for the combination? You said cost savings and what else?
`MS. LUTTON: And redundancy, and I just cited some
`testimony from Dr. Horst referring to redundancy and each of the
`references talks about redundancy being one of the goals. These
`are fault tolerant redundant systems. That's the entire purpose
`behind the disclosures and behind the designs.
`JUDGE QUINN: So with regards to redundancy, I
`think one of my colleagues already asked that Mylex already has
`redundancy, so what extra redundancy are you advocating here
`that you have evidence of?
`MS. LUTTON: So the evidence is from Hathorn itself,
`which shows a system that has a single -- you know, has a direct
`communication path and Hathorn is proposing moving to this
`new communication structure for redundancy purposes. So one
`of skill in the art looking at that would be motivated to try that
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 22
`
`
`
`
`
`
`
`Cases IPR2014-00901 and IPR2014-00949
`Patent 6,978,346 B2
`
`combination and apply that to Mylex to see if they can achieve
`the same redundancy considerations in Hathorn and that includes
`over distance, because both Mylex and Hathorn talk about
`distance.
`JUDGE QUINN: Where in Hathorn do you precisely
`say that the failover was done for redundancy? Because I only
`see here that it was a reduction in links, not that it was a
`purposeful intentional adding of links