`U.S. PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`CISCO SYSTEMS, INC.
`Petitioner,
`
`v.
`
`EGENERA, INC.
`Patent Owner.
`____________
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`____________
`
`Record of Oral Hearing Held:
`July 25, 2018
`____________
`
`Before KRISTEN L. DROESCH, CHARLES J. BOUDREAU and
`MELISSA A. HAAPALA, Administrative Patent Judges.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`THEODORE M. FOSTER, ESQUIRE
`DAVID L. MCCOMBS, ESQUIRE
`Haynes and Boone LLP
`2323 Victory Avenue, Suite 700
`Dallas, Texas 75219
`
`JAMES E. QUIGLEY, ESQUIRE
`JOHN B. CAMPBELL, ESQUIRE
`McKool Smith
`300 West 6th Street, Suite 1700
`Austin, Texas 78701
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`
`
`
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`
`
`
`
`
`
`The above-entitled matter came up for hearing on Wednesday, July
`
`25, 2018, commencing at 12:59 p.m., at the U.S. Patent and Trademark
`Office, 600 Dulany Street, Alexandria, Virginia.
`
`
`
`
`
`
`
`2
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`
`P R O C E E D I N G S
`- - - - -
`JUDGE DROESCH: Please be seated. Good afternoon. We're here for the
`oral hearing for IPR 2017-01340 between Petitioner, Cisco Systems and
`Egenera. I'm Judge Droesch. We're joined by video with Judge Haapala
`from the Denver satellite office and Judge Boudreau from the San Jose
`satellite office. Each party has 60 minutes total time for arguments.
`Because Petitioner bears the burden of showing unpatentability, Petitioner
`will open the hearing by presenting its arguments first. Petitioner may also
`reserve some of its time for rebuttal; and following Petitioner's arguments,
`Patent Owner will proceed; and then, if applicable, Petitioner may present its
`rebuttal arguments.
`
`Counsel for Petitioner, whenever you're ready, please state your name
`for the record and introduce your co-counsel and anyone else in attendance
`today; and then you may begin your arguments after the introductions.
`
`MR. MCCOMBS: Good afternoon, Your Honors. I'm David
`McCombs here on behalf of Petitioner, Cisco Systems. With me is my
`partner, Theodore Foster, who will be making the presentation today; and
`also, we have one of Cisco's attorneys, Peter Magic, in attendance with us.
`Thank you.
`
`JUDGE DROESCH: Thanks.
`
`MR. MCCOMBS: I'll turn it over to Theo.
`
`JUDGE DROESCH: Okay.
`
`MR. FOSTER: Good afternoon. May it please the Board, Theo
`Foster on behalf of Petitioner, Cisco Systems. Looking at slide 3, there are a
`
`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
`
`
`
`3
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`few topics that I'd like to discuss here this afternoon, and the first of those is
`Egenera's attempt at swearing behind the Ma reference, and their failure to
`do so. As we know, swearing behind requires both conception and
`diligence, and Egenera's evidence fails on both counts. I think this case is
`somewhat notable. There are declarations from two named inventors,
`neither of them in their declarations claims to have conceived of the subject
`matter prior to Ma's filing date; and neither of them claims to have been
`diligent in pursuing those ideas up to the filing date of their non-provisional
`application. So for that reason alone, the fact that the named inventors
`themselves don't even claim to have conceived of the claimed concepts or
`been diligent in pursing them, for that reason alone, this swear-behind
`attempt would fail.
`
`But moving on and looking at slide 4 -- and looking, specifically, at
`some of Egenera's evidence presented on this topic -- slide 4 shows a portion
`of the Patent Owner's Response at page 19 where they address their alleged
`conception evidence relating to the failover logic limitation. In particular,
`the logic to assign the virtual MAC address of the failed processor to the
`processor that replaces the failed processor.
`
`And on this point, Egenera says well, back above in the earlier
`analysis for portion 1(c) they allege that they showed that their node
`configuration includes a virtual MAC address; but if we move to slide 5 --
`and slide 5 here has the Patent Owner's Response at page 11 where they
`address what they label as limitation 1(c) -- and their evidence does not
`support their allegation that the inventors -- the named inventors -- had
`conceived of the virtual MAC address as being part of a node configuration.
`
`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
`
`
`
`4
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`Instead, what their evidence says is that a Giganet driver code will assign an
`Ethernet 48-bit media access control, or MAC address, in place of a
`hardware-provided MAC address; but it's not described as being part of a
`node configuration.
`
`And if we move on and look at slide 6, the evidence that Egenera
`provided from what they labeled their corroboration documents, actually
`shows that this idea of a MAC address being moved upon failover from
`machine-to-machine, that it appears the named inventors had not yet
`conceived of that idea prior to Ma's filing date. Looking at slide 6 -- and this
`is in the middle -- there's a portion of Egenera's Exhibit 2012, from page 6,
`where there's discussion of the format of a simulated MAC address -- and it
`says that's to be specified; so, obviously, they were still figuring out how
`they were going to handle MAC addresses, if at all. But it goes on and it
`says that the simulated-MAC address "will include the node number." So,
`the node number, the number of a node, is particular to a MAC address for
`that node. Obviously, this implies that any simulated MAC address that they
`were conceiving of was going to be specific to a particular node, not
`something that could be transitioned, as the claim calls for, from one
`computer processor to another computer processor.
`
`So, the conception evidence is simply not here and, in fact, what
`Egenera has put forward, affirmatively shows that they had not yet come
`upon the idea of virtual MAC addresses that would be re-assignable between
`nodes or between computer processors. Even if Egenera had evidence of
`conception for all of the claim limitations -- and, I think, I've just
`
`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
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`demonstrated they don't -- even if they had that, of course, they also have to
`show diligence.
`
`And if we look at slide 7, Egenera's evidence regarding diligence is
`wanting as well. In their Patent Owner's Response, the attorneys assert that
`Egenera was working on the claimed subject matter from November 2000
`until January 2002 -- and I believe that's in reference to the January 4, 2002
`filing date of their non-provisional application. But the declarations from
`the named inventors don't support that allegation from the attorneys; and, in
`fact, the declaration of named inventor, Scott Geng, merely states that the
`group was working "during the late 2000 time period and into 2001." So,
`there's not even a claim or a statement to support the allegation of work into
`2002. It was merely from an allegation of work from late-2000 until
`sometime in 2001, but ending, apparently, well before the January 2002
`filing date of the non-provisional application. So, the diligence aspect of
`this swear-behind attempt is also lacking.
`
`With that, I believe Ma remains prior art, and I'd like to move on
`looking at slide 8, to the second topic, and that's the obviousness of
`combining the Aziz and Ma references.
`
`If we go to slide 9, I have here a quote from a treatise, just
`background treatise, on high availability systems and servers failover
`concepts -- the general technology space that we're dealing with here of
`network computing -- and Marcus is Exhibit 1009 -- we filed a second copy
`as Exhibit 1024 in response to an objection from Egenera. Pages 88 and 89
`of those exhibits contain background information that a person of ordinary
`skill in the art would have known having worked and been familiar with
`
`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
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`server systems, been familiar with the failover concepts that were out there
`and were known in the industry; and what Marcus describes is a known
`problem that can occur when a server fails over from a failed node to a
`replacement, and that is that the clients that were accessing that server will
`have cashed a mapping between the IP address and the MAC address of that
`server that has now failed; and that mapping can become out of date or stale
`incorrect depending on how that failover happens.
`
`If the client continues to rely on that stale information, of course, they
`will continue to send messages that won't reach the replacement server if the
`replacement server has a different IP address or MAC address. And so, this
`problem was known; and as Marcus demonstrates, the solution to it was also
`known. As we see at the bottom of slide 9, one solution is to move the
`MAC address, as well -- move the MAC address from the primary to the
`secondary node; move it from the failed computer to the replacement
`computer.
`
`So, Marcus is describing this general background information and
`knowledge about how failover works, how networks work, and how servers
`and clients interoperate that a person of ordinary skill in the art would have
`brought with them to their understanding of Aziz, and their understanding of
`Aziz's limitations -- and, in particular, that Aziz being a system that is
`performing failover -- Aziz's system is susceptible to this stale IP/MAC
`address mapping problem, and Ma, of course, is our secondary reference that
`teaches the known solution to that; and, in particular, Ma is providing that
`solution in a virtualized network context. And so, since Aziz is describing
`virtual networking and would suffer from this known problem, we believe it
`
`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
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`would be obvious -- and our expert, Dr. Shenoy, supported us on this -- that
`it would have been obvious to apply the known solution of Ma -- of
`assigning IP and MAC addresses together to a replacement server during a
`failover operation. And so, that combination then comes together with Aziz
`having most of the claim limitations and then Ma coming in to teach this
`known solution to a known problem that a person of ordinary skill in the art
`would have recognized Aziz suffers from.
`
`I'd like to move on to slide 13 and the next topic; and that is Aziz's
`teaching about failover and the automation of the failover operation.
`
`JUDGE HAAPALA: Counselor, before we move on, going back to
`Ma's teaching, the virtual MAC address on the active commander, the active
`commander is a switch and not a processor, is that right?
`
`MR. FOSTER: So, looking at slide 11, I believe might be helpful for
`this question. So, Ma's teaching is that, yes and what it terms as an active
`commander -- well the commander activator is a separate thing -- the active
`commander is the device that is in ownership or possession of the IP and
`MAC addresses -- it's the device that's functioning in the role and if that
`device fails then Ma teaches assigning those addresses to a replacement
`device. Now, in Ma -- Ma describes that in the context of what Ma terms a
`network device -- Ma's specific example is, as you suggested in the context
`possibly of a switch or a router, but Ma is not limited to that; and if we move
`to slide 20, Ma's teaching is more general than that and its concept of what a
`network device could be is much broader than just switches, bridges, and
`routers as Egenera is arguing.
`
`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
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`Looking at the second quote on slide 20 -- this is from Ma, column 2,
`
`starting around line 29 -- Ma refers to "hosts or other network devices"
`plainly indicating that hosts are also within the concept of network devices
`within Ma. And if we look at figure 1(a) of Ma, also on slide 20, we can see
`that the host 5(a), 5(b), and 5(k) are depicted as ordinary computers; and so,
`the teachings of Ma are not restricted simply to switches but, in general, to
`all kinds of devices found in a network. Judge Haapala, does that answer
`your question?
`
`JUDGE HAAPALA: Yes, thank you.
`
`MR. FOSTER: Thank you. So, moving back to slide 14 and Aziz's
`teaching of automated failover, Aziz describes this in a couple of different
`ways. The first I have here a quote from Aziz, column 7, line 8 where Aziz
`talks about what it's going to do with its computing elements; and it
`describes that it's going to have a pool of unused or idle computing elements,
`and that those can be assigned to a Virtual Server Farm, a VSF in Aziz's
`terminology -- which is basically a virtual network. An idle element can be
`assigned to a Virtual Server Farm "to deal with failures of a particular
`computing element in a VSF."
`
`So, Aziz is contemplating a failover situation and a failover capability
`with this language; and if we continue on, looking at Aziz column 3, starting
`around line 17, Aziz describes the process that it's going to use and the
`technology it's going to use for moving those computing elements from the
`idle pool into a VSF, and it describes doing that "under program control."
`So, this is the teaching that -- this process of employing or deploying a spare
`computing element to replace a failed computing element -- that's done
`
`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
`
`
`
`9
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`under program control; and Aziz emphasizes the automated nature of this as
`we see in column 8 around line 66, and then spilling over to the top of
`column 9, Aziz states that "a particular VSF, or Virtual Server Farm, is not
`subjected to manual reconfiguration." So, Aziz is very specific that its
`process for managing Virtual Server Farms is not manual; it is automated.
`
`And if we move on and look at slide 15, Aziz actually disparages
`manual processes repeatedly. Looking at column 2 of Aziz, line 49, Aziz
`refers to how human error in configuring a new server can cause the entire
`server farm to malfunction; so, obviously, a bad result. In column 1, line 50
`it refers to prior art solutions that suffered from "manual and error-prone
`administrative steps." And then at the end of Aziz, it touts its advantages
`over conventional manual approaches to constructing and managing server
`farms. So, Aziz is very critical of manual processes, and is emphasizing
`with this language the automated nature of its solution, and the fact that by
`deploying idle elements to deal with failures under program control, Aziz is
`teaching automated failover techniques.
`
`The next topic that I'd planned to talk about was the teaching of
`virtual MAC addresses for computer processors; and Judge Haapala, I think
`in answering your earlier question, I addressed most of what I was going to
`say on this. I would just point out that Egenera's arguments relating to this
`concept --
`
`JUDGE HAAPALA: Counselor, are you on a particular slide?
`
`MR. FOSTER: Yes; I'm sorry, I'm on slide 19.
`
`JUDGE HAAPALA: Thank you.
`
`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
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`MR. FOSTER: Egenera's arguments relating to the claimed virtual
`
`MAC addresses for computer processors. Egenera's arguments on that rely
`entirely on their misunderstanding or misinterpretation of the Ma reference
`in their assertion that Ma is somehow limited to only switches, bridges, and
`routers when, as I've discussed earlier, in fact, it's very clear that it's more
`broad than that. I would also point out that Ma, at column 2, line 40, Ma
`refers to a host or client, and then in parentheses, it says, i.e., the network
`device. So it's renaming network device and indicating by network device, it
`also means hosts and clients. It's very clear that it's not just switches,
`bridges, and routers.
`
`Looking a slide 24, the final topic that I wanted to address this
`afternoon relates to the claimed construction for Virtual Local Area
`Communication Network. And looking at slide 25, the Board at institution
`addressed this limitation -- this phrase -- and was not persuaded by Egenera's
`arguments then that the term should be interpreted as requiring the entire
`network be simulated by software. The board rejected that view -- noting
`that the '044 specification refers to the hardware components and the
`physical network underlying the claimed technology. And looking a slide
`26, Egenera has simply repackaged this argument and dressed it up in some
`new clothes, but fundamentally, I think it's still the same argument. Egenera
`in their Patent Owner Response, at page 34 -- this is again, on slide 26 --
`Egenera is now arguing that this Virtual Local Area Communication
`Network should be understood to require using software to implement
`network equipment itself. So, this is simply repackaging this argument 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
`
`
`
`11
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`the entire network itself is software implemented, and that's, again,
`inconsistent with the '044 specification.
`
`Looking at slide 26 and the '044 specification at column 3, around line
`21, it refers to the two switch fabrics which are the network components and
`the two switch fabrics being contained in a single chassis, strongly
`indicating the physical nature of those; and then at column 3, line 57, it goes
`on and describes the switch fabrics as including devices such as a NIC-
`CLAN 1000 and a CLAN 5300 switch; and we asked Peter Manca, one of
`the named inventors about those devices -- what are the NIC-CLAN 100 and
`the CLAN 5300 -- and in his deposition, this is Exhibit 1034, page 39,
`starting at line 15 -- he confirmed that those are, indeed, physical
`components. So, Egenera's argument that it's the network equipment itself
`that is being software simulated is inconsistent with the disclosure of the
`'044 specification, and I would encourage the Board to reaffirm the initial
`construction for Virtual Local Area Communication Network is merely
`requiring partial simulation; and to that idea and the partial simulation
`looking at slide 28, either under the Board's initial construction or even
`under Egenera's proposed construction, Aziz is very clear that it does teach
`software simulated network equipment.
`
`For example, looking at Aziz's figure 3, you can see that element 302
`is a computing element that has been configured in a Virtual Server Farm to
`act as a load balancer in a firewall; and Aziz describes how that load
`balancer and firewall operate. For example, looking at column 18, starting
`at line 16, Aziz describes how all of the computing elements -- they're just
`computer processors with network and storage connectivity -- and as Aziz
`
`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
`
`
`
`12
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`says, they're fungible, right; they're completely interchangeable and inter-
`replaceable. They only assume a specific processing role when they are
`connected to a particular Virtual Server Farm and they load the software
`that's been assigned to them. And so, it's the software that is making a
`computing element act as a firewall in Aziz in figure 3.
`
`In Aziz, at column 9, lines 52-54, it reiterates that, for example, a
`CPU can take on a particular role by be pointed to and then loading
`"dedicated load balancing/firewalling software." So, here we have Aziz
`describing network equipment such as a firewall or a load balancer that's
`being simulated or emulated, being assigned to that role through the use of
`software. So, here's our partial simulation of network equipment.
`If there are no further questions from the Board --
`JUDGE HAAPALA: I have nothing further.
`
`MR. FOSTER: Okay; very good.
`
`JUDGE DROESCH: I have nothing.
`
`MR. FOSTER: I encourage the Board to reaffirm its initial findings
`
`and find all of the claim limitations taught by the prior art and the claims
`obvious and unpatentable as put forward in the Petition. I'll reserve the
`remainder of my time for rebuttal.
`
`JUDGE DROESCH: Thank you.
`
`MR. FOSTER: Thank you.
`
`JUDGE DROESCH: Whenever you're ready to begin, please
`introduce yourself and your co-counsel, and then whenever you're ready,
`begin.
`
`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
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`MR. QUIGLEY: Good afternoon, Your Honors. James Quigley of
`
`McKool Smith here on behalf of Egenera, Inc. Also with me is John
`Campbell, also with McKool Smith.
`
`This is not a case with allegations of anticipation; this is not a case
`where Cisco is suggesting that there's an individual reference that's teaching
`every limitation in the '044 Patent; instead this is a case of obviousness. As
`we see here on slide 2, in this case, Cisco is suggesting that the Ma reference
`in combination with Aziz renders obvious the independent claims of the '044
`Patent, claims 1 and 4; but, in fact, that combination of Ma with Aziz fails to
`render the claims of the '044 Patent obvious.
`
`Slide 3 has the three independent reasons why Cisco's invalidity or
`obviousness challenge fails here in this case. First reason -- the '044 Patent
`pre-dates Ma. The '044 Patent and its claimed inventions were conceived of
`more than a month before the Ma reference was filed for. The '044 Patent
`inventions were diligently reduced to practice and Ma's not prior art. If Ma's
`not prior art, Cisco's entire challenge fails.
`
`The second independent reason that Cisco's obviousness challenge
`fails is that the combination of Ma with Aziz fails to render obvious the
`claimed virtual MAC addresses for the claimed computer processors.
`Instead, what we have with Cisco's Ma reference is a reference that refers to
`"virtual MAC addresses" in the context of network devices -- switches,
`bridges, and routers -- and not the claimed computer processors. In fact --
`
`JUDGE HAAPALA: Counselor, even if that's true, I'm going to
`interrupt you here -- even if that's true, is there a difference in how a switch
`or router would use a MAC address, virtual MAC address, as opposed to a
`
`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
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`processor? Isn't the virtual MAC address just an address used to
`communicate? Is there some difference with the virtual MAC being
`assigned to a switch or a router?
`
`MR. QUIGLEY: Yes, I do, Judge Haapala, think that there is a
`difference between Ma's "virtual MAC address" that's being used for these
`network devices and then the MAC addresses that are being used in the '044
`Patent. And the difference here is that the Ma patent is addressing a
`different problem with these network devices than with a computer
`processor in the context of the '044 Patent; and it's addressing that problem
`in a different way, with a different sort of solution.
`
`JUDGE HAAPALA: Okay; Counsel, that's not what I'm asking. I'm
`asking specifically about the virtual MAC address and if there's a difference
`on how a switch or router uses a virtual MAC address and a processor uses a
`virtual MAC address.
`
`MR. QUIGLEY: I think, in the context of this case, with these
`network devices in Ma, like Ma’s switches, yes there is a difference in how
`the MAC address is being used. Ma is dealing with a virtual MAC address
`for an entire group of devices; it's not even assigned to an individual device.
`
`JUDGE HAAPALA: My understanding though is that it's being used
`to communicate with other devices, correct? The active commander's virtual
`MAC address is being used to communicate with other devices?
`
`MR. QUIGLEY: In some way it's involved in the communication
`process in Ma, yes.
`
`JUDGE HAAPALA: Okay; thank you.
`
`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
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`MR. QUIGLEY: And so, the third independent reason that the '044
`
`Patent is not rendered obvious by Cisco's prior art is that the combination of
`Ma with Aziz fails to render obvious the claim to automated failover.
`
`Moving on to slide 4, the first independent basis that Cisco's
`obviousness challenge fails is that Ma is not prior art. The '044 Patent was
`conceived of before Ma; it was diligent, reduced to practice; and without
`Ma's prior art Cisco's challenge fails.
`
`We see on slide 5, that Ma has a filing date of December 15, 2000; but
`moving on to slide 6, the '044 Patent and its inventions were conceived of by
`November 7, 2000. How do we know this? Well, we've actually been able
`to go back and get internal Egenera documents from back in the 2000 time
`period. These internal Egenera documents related to a product that the
`inventors were working on called the Interframe. I'm almost certain you
`hadn't heard of the Interframe before this case -- it was the internal name for
`a product that later became as the BladeFrame when it was released -- but
`these documents related to the Interframe show that the inventors had
`conceived of each and every limitation in the '044 Patent by November 7,
`2000.
`Dr. Jonathan Chao, an engineering professor at NYU, has looked at
`
`these documents and he's told us that those documents show the support for
`the claims of the '044 Patent. On the other hand, Cisco has offered no expert
`testimony related to these conception documents. In terms of the dates of
`these documents -- of these conception documents -- we're not relying solely
`on inventor recollection; we're not relying solely on the dates listed on the
`documents; and so we've actually been able to go back into the repositories
`
`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
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`where these documents have been housed for a number of years and look at
`the metadata; and the metadata has told us that each of these documents was
`last modified on or before November 7, 2000.
`
`So, in this case, Cisco really only disputed in its Reply Brief two
`limitations -- whether or not two limitations in each of the independent
`claims were supported by these conception documents. The first one, as we
`see on slide 6, is this configuration logic limitation. Cisco didn't address that
`configuration logic limitation this afternoon. I don't plan to address it unless
`any of Your Honors would like hear about it.
`
`JUDGE HAAPALA: I'm more interested in the failover logic myself.
`
`MR. QUIGLEY: Yes, Your Honor. So, that second limitation is here
`on slide 6 -- it's that failover logic limitation; and Cisco is contesting
`whether or not this failover logic limitation is shown in these conception
`documents. So before we jump into conception documents, I want to give
`you a bit of background on the Interframe. So, I'm turning now to slide 9.
`
`So, like I said, the Interframe is something you probably hadn't heard
`of by 2000. It was an internal name for a product that ended up being
`released as the BladeFrame. The Interframe was a computer processing
`platform. It was a platform that was connectible to both an external storage
`and external communication network; and we can see here on slide 9 an
`annotated version of figure 1 showing us how this grey Interframe platform
`could be connected to this yellow and blue communication and storage
`network.
`
`In slide 10, I've actually peeled away that grey highlighting for the
`Interframe platform to show you what's underneath in some of those relevant
`
`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
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`components that are going to be discussed this afternoon with respect to
`failover logic. On the left-hand side of slide 10, we have in purple the
`processing or application nodes. Sometimes these are referred to as
`processing blades in these conception documents. These were the compute
`resources that are a part of this Interframe platform.
`
`On the middle right-hand side of slide 10, we have in red the
`Interframe controller node. This Interframe controller node was how the
`Interframe platform was able to communicate with these external and
`communication networks; and in the middle we see in green the Giganet
`internal network. The Giganet internal network within the Interframe is
`what allowed those processing application nodes to communicate with that
`Interframe controller node.
`
`Jumping ahead to slide 14 and that failover logic limitation that we
`were just discussing -- that limitation reads, the failover logic including logic
`to assign the virtual MAC address of the failed processor to the processor
`that replaces the failed processor. We see here on slide 14 from Egenera's
`Patent Owner Response, a description of how the Interframe platform
`included failover logic such as on the Interframe controller node that we just
`highlighted in red, which when a processing blade or node failed -- those
`were highlighted in purple -- caused another blade or node to replace it,
`which included taking on the failed blade or node's configuration and
`resources. Part of those resources included a virtual MAC address. Now the
`documents actually bear this out. The documents bear out that there's an
`automatic failover process that was planned for this Egenera Interframe
`product.
`
`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
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`Now, moving on to slide 16, we've got to take a step back with this
`
`failover logic and remember that the failover logic in the claims is
`addressing when there's a failure, we've got to allocate a new computer
`processor to take the place of a failed processor. We've got to do that
`automatically. We see on the left-hand side of slide 16, from Exhibit 2024, a
`document entitled Egenera Interframe Architecture, a quote that reads "the
`health of an application node is continuously assessed by management
`software through heartbeats. Management software can notice when an
`application node has crashed or hung and can initiate recovery behavior."
`What this is telling us is the software on the Interframe controller node is
`continuously monitoring the health of these processing application nodes.
`It's doing that through this heartbeat mechanism; and it's able to determine
`when there's been a crash, a hang, or some type of failure; and when that
`failure occurs, it's able to take some pre-planned, pre-specified route.
`
`On the right-hand side of slide 16, also from Exhibit 2024, we see that
`the application nodes have configurable behavior in response to certain
`events such as that failure we were just talking about. The response can be
`something like automatically booting a new processing node with a specified
`configuration. Now, I don't have these documents up on a slide, and I don't
`have the ELMO here -- I'm sorry Judge Droesch -- I'm just not connected in
`a way that I can show you -- but Exhibit 2009 and 2010 recited in Patent
`Owner's Response, pages 15-17 -- and there are some relevant excerpts to
`this automatic failover, this automatic replacement of a failure processing
`node.
`
`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
`
`
`
`Appeal IPR 2017-01340
`Patent 6,971,044 B2
`
`
`Exhibit 2009, at the bottom of page 2 in the left-hand column reads
`
`spare application processors can be configured as standby nodes, and can be
`rapidly booted with the same software