ADAPTIVE SPECTRUM AND SIGNAL ALIGNMENT INC v BRITISH TELECOMMUNICATIONS PLC

This is my decision on a Part 8 claim for declaratory relief in respect of the proper interpretation of a clause in a patent licence dated 27 July 2015 (the “Licence”) entered into between the Claimant, Adaptive Spectrum and Signal Alignment Inc (“ASSIA”) and the Defendant, British Telecommunications plc (“BT”). Under the Licence each party granted the other a licence in respect of certain patents, the relevant licence being that granted by ASSIA to BT. Also relevant is a settlement agreement and release entered into by the parties on the same date (the “Settlement Agreement”).

ASSIA is a California incorporated corporation that was founded by the inventor Professor John Cioffi. It holds a number of patents in the field of Digital Subscriber Line (“DSL”) technology, which it has licensed to a number of communications providers (“CPs”). Until May 2022 ASSIA’s primary business was the provision of technology, including software solutions, to CPs. It has now sold that business but retains most of its patent portfolio, including patents relevant to this dispute.

DSL technology is the technology that enables copper wires originally intended for voice telephony to support high volumes of data and deliver broadband internet access, using higher frequencies than those used for voice communication. DSL technology requires there to be a modem at each end of the copper line. It is relevant where premises are connected to local exchanges entirely via copper wires and (as is now more often the case) where the copper line is used only between the premises and the relevant street cabinet, with the distance between the street cabinet and the exchange being covered by optical fibre (Fibre to the Cabinet, or “FTTC”). FTTC allows for faster data transmission, simply because a smaller distance is dependent on copper wiring.

The particular subject matter of the dispute is Dynamic Line Management technology (“DLM”). This is technology that monitors the performance of DSL lines and, based on that performance, applies re-profiling to improve performance. ASSIA’s products included a product called DSL Expresse®, which performs DLM functions.

BT, through its division (now subsidiary) Openreach, controls most of the UK’s telecommunications network infrastructure. Openreach sells access to its network and/or more extensive services both to third party CPs and internally to other divisions of BT. Those divisions comprise BT’s own CP in the form of BT’s retail division, now called BT Consumer, and a wholesale division now called BT Enterprise. BT Enterprise supplies both certain third party CPs and BT Consumer. BT also holds a number of patents, including in respect of DSL technology. BT’s own DLM controllers are known as Rate Adaptive Monitoring Boxes, or “RAMBos”.

For ease of reference, I will use the current names of the various BT divisions and also refer to persons to whom CPs make supplies (who may be consumers or businesses) as “end users”.

The Licence was entered into as part of the settlement terms of a dispute which had been initiated in 2011 by ASSIA commencing proceedings against BT in this jurisdiction for infringement of certain patents, but which was subsequently expanded by each party commencing opposition proceedings in the European Patent Office (EPO) and by BT commencing its own infringement proceedings against ASSIA and one of ASSIA’s customers in the Delaware District Court.

At the point of settlement, BT had been found by the Court of Appeal, on appeal from Birss J, to have infringed ASSIA’s patents in respect of one version of its RAMBos, and BT had made two different modifications that it claimed addressed the infringement. Birss J had rejected BT’s claim for a declaration of non-infringement in respect of the first modified version (albeit that decision was subject to appeal). However, BT was seeking summary judgment in respect of the second modified version, and expressed confidence in it on the basis that it used a mechanism employed in BT’s other networks that ASSIA was not challenging. In the meantime, a damages inquiry had been initiated by ASSIA. The relevant reported decisions are ASSIA v BT [2013] EWHC 3768 (Pat); [2014] EWCA Civ 1462, and (for Birss J’s decision in respect of the first modification) ASSIA v BT No 2 [2014] EWHC 4194 (Pat).

There were several elements to the overall settlement, which are discussed further below. Broadly, each of BT and ASSIA granted a royalty free licence to the other in respect of their relevant patents and agreed to discontinue the various proceedings. BT also agreed to make certain payments to ASSIA and ASSIA agreed to provide certain public support and assistance to BT. They agreed a mutual covenant not to sue.

There are three preliminary points. First, both parties asserted before me that they were in a strong negotiating position in respect of the settlement. I am not in a position to assess the strength of their respective positions, beyond the findings that I have just set out as to the history of the dispute and the fact that they reached agreement on the terms that they did. In particular, I am not in a position to assess the value of the respective licences granted by the parties, or to compare the monetary amounts that BT agreed to pay with the value to BT of the support and assistance that ASSIA committed to provide.

The second preliminary point relates to confidentiality. Some of the evidence that I was taken to is commercially sensitive. From BT’s perspective this includes in particular the quantum and breakdown of the monetary consideration it provided to ASSIA under the terms of the settlement. I accepted that there was sensitivity and made an interim order at the trial to preserve confidentiality, notwithstanding that the relevant documents were referred to and read by the court. However, I also requested written submissions about the approach that should be taken in this judgment in relation to the detail of the monetary consideration, having regard to the requirement for open justice. I am grateful for those submissions.

In the result, and applying the principles in Dring v Cape Intermediate Holdings [2019] UKSC 38; [2020] AC 629, I have concluded that there is no need to refer to the amount or precise details of the monetary consideration in order for the reasons for my decision to be fully understood. Taking account of the point made at [10] above, those details are simply not material to my reasoning. I have therefore not included them. As regards both that information and other commercially sensitive information identified by BT, namely sums that BT has obtained from licensing certain patents, I propose to make an order that will ensure that third parties are prevented from obtaining copies of the relevant documents, or at least unredacted versions of them, without the approval of the court. I will also invite submissions to determine whether there should be any continuing order to prevent the collateral use of confidential information by the parties.

The third preliminary point to make is that both Mr Segan for ASSIA and Mr Saunders for BT rightly proceeded on the basis that a substantial amount of the evidence that had been served by both parties was inadmissible, because it concerned the parties’ subjective intentions or related to the negotiations that led up to the entry into of the Licence and the Settlement Agreement. For present purposes it is not necessary to attribute blame for that state of affairs, but I should record that I have reached my conclusions by reference to the other factual evidence and by reference to the submissions received, excluding consideration of the offending material.

BT has three different DSL networks, known as 20CN, 21CN and NGA (referring to the 20th Century, 21st Century, and “Next Generation Access” networks respectively). The 20CN and 21CN networks use versions of DSL technology called ADSL (Asymmetric Digital Subscriber Technology), and operate where FTTC is not available or is not used by the customer in question. The NGA network operates where FTTC is in use. The majority of lines in the NGA network use a different type of ADSL technology called VDSL2, and the remainder use a type called G.Fast, which started to be introduced in around 2018 (following a pilot in 2017) and which facilitates ultrafast broadband. BT operates DLM controls via a single RAMBo for each of the three DSL networks. In the previous English proceedings it was the NGA RAMBo that was found to infringe two of ASSIA’s patents.

As already mentioned, for DSL technology to work there must be a modem at each end of the copper line. The modem at the customer’s premises is referred to as the Customer Premises Equipment (“CPE”). Whilst it can be provided separately from the customer’s router (a “two box” solution) it is more commonly provided in a single box with the router. (The router is what allows the customer to connect multiple devices, either wirelessly or via ethernet, to the internet.) In its “one box” format, CPE is typically provided by the CP with which the end user contracts.

BT’s evidence, which I accept, was that BT has used the two box solution when introducing new technology, in case unforeseen problems arise, before transitioning to a single box. The one box solution is also known as “wires only”, because from the CP’s perspective no other physical item is provided by BT at the customer end (apart from a socket). In contrast, where a two box solution is provided the modem “box” is provided by BT rather than by the CP, and the router is generally provided by the CP. There are no material technical differences between the two box and one box alternatives, but the latter is neater for the end user and facilitates the CP’s brand recognition, because it can design and apply its branding to the single box.

The modem at the exchange or street cabinet end of the copper line will typically be one of many comprised in a piece of equipment known as a Digital Subscriber Line Access Multiplexer or “DSLAM”. The DSLAM aggregates internet-bound communications from the various modems it contains for onward communication on optical fibre, and disaggregates customer-bound communications to each modem for onward transmission in the other direction along each copper line.

The two modems at each end of the copper line communicate via a shared communications protocol, which is achieved via an automatic negotiation process between them called initialisation. Once that is complete the normal operation process, called “SHOWTIME”, commences.

On the network side of the DSLAM, connection to the internet is via a series of different optical fibre connections through different exchanges. Part of this includes each CP’s “backhaul”, a term used to refer to a CP’s own connection between the infrastructure nearer the customer and the internet.

In broad terms BT’s DLM technology operates as follows. Each line uses one of a number of predefined “profiles” which constrains how the modems communicate. Once a day each DSLAM reports certain performance data to the relevant RAMBo, which the RAMBo evaluates. This may result in the RAMBo issuing a command to the DSLAM requiring a change in profile for a particular line. Where it does so, the DSLAM reacts by requiring the relevant pair of modems to “re-train” using the new profile. The DSLAM does not itself perform DLM, rather it responds to instructions given to it. The CPE similarly has no DLM functionality, but reacts to the instruction given.

BT’s relevant arrangements with CPs fall into two different categories:

What BT refers to as an “unbundled” or Metallic Path Facility (“MPF”) product, which involves BT leasing the relevant copper line, or rather access to it, and space in the exchange or street cabinet to a CP, with the CP installing its own DSLAMs and being responsible for any DLM. Where the copper wire connection is to an exchange (that is, the 20CN and 21CN networks) this is referred to as Local Loop Unbundling or “LLU”. Where FTTC is used so that the connection is to a DSLAM in a street cabinet (that is, the NGA network), it is referred to as Sub-Loop Unbundling or “SLU”. In the latter case the CP would also usually have to provide its own fibre between the street cabinet and the exchange, as well as providing DSLAMs in each street cabinet.

In a “bundled” product BT provides a data connection rather than mere access to physical infrastructure, and does so all the way between the customer premises and a hand-over point in the network where the CP may connect their backhaul (the “headend”). This would involve using a BT DSLAM either in the exchange or (for FTTC) the street cabinet. This case concerns the NGA network where FTTC is in use, so the DSLAM will be in a street cabinet. However, the hand-over point to the CP will be in the exchange, so the data connection that BT provides will run partly over copper wires and partly over optical fibre. Unlike the MPF product, which is limited to access to the physical infrastructure, the data connection provided by the bundled product also incorporates a connection at the ethernet “layer”. This arrangement is sometimes referred to as Virtual Unbundled Local Access or “VULA”. (The reference to unbundling in the term VULA initially appears confusing given that it is described as a bundled product. The explanation is that the reference to unbundled here is a more generic one, referring to the regulatory unbundling of telecoms infrastructure to provide a competitive market.)

By way of brief explanation of the reference to a “layer”, the provision of an internet service requires a number of layers of functionality. The physical infrastructure over which electrical signals are transmitted is the bottom layer. The ethernet layer is at level 2. The user level layer is layer 7. (In fact, VULA also incorporates one element of layer 3, known as “multicast”, but this is not material for present purposes.)

A CP purchasing an unbundled product from BT is, as already indicated, responsible for its own DLM. It may choose to buy an ASSIA product for that purpose, in particular DSL Expresse®. In contrast, a purchaser of VULA has no control of the relevant DSLAMs and therefore has no ability to apply its own DLM. Instead, where VULA is operated BT both owns and operates the relevant DSLAMs, and uses its RAMBos to control performance of the DSL communication, just as it does in respect of lines allocated to BT’s own customers. Individual DSLAMs may carry communications for a number of different CPs, including of course BT Consumer.

A CP may hold a mixture of bundled and unbundled products for different lines. Sky UK Limited (“Sky”) is an example of this, having previously bought ASSIA’s DSL Expresse® product for unbundled lines.

Where BT provides the VULA product to CPs, it enters into contractual arrangements with them on standard published terms and conditions referred to as the “GEA Conditions”. (GEA, or Generic Ethernet Access, or strictly its variant GEA-FTTC, is BT’s term for VULA as used in its NGA network.) The version of the GEA Conditions included in the documentary evidence dates from 2019, but there was no indication that material relevant changes had been made since 2015.

As far as end users are concerned, they contract only with CPs (which again may include BT Consumer). There is no indication that the contractual arrangements with any CP differ according to the type of DSL product held by the CP. Further, while CPE is typically provided by the CP, it is not obligatory for end users to use it in order for the connection to work. BT’s evidence was that modem functionality is standardised and end users could provide their own. Further, some small CPs buy their CPE from BT. ASSIA accepted that modem technology is standardised. However, Mr Segan pointed out that there was no evidence about consumers actually using CPE not provided by the relevant CP, and suggested that CPE sourced from one CP could require reprogramming to work with another CP’s service.

FTTC was first introduced in around 2009. Although by the end of 2014 around 60% of street cabinets had physical FTTC infrastructure, customer migration was more gradual. In early 2015 there were around 4 million lines using VDSL2 and connected to an FTTC service. However, since 2015 the NGA network has grown to approximately 14.9 million lines using VDSL2, driven by increased customer uptake of fibre broadband products. This is now combined with what is currently a much smaller number using G.Fast.

The unbundled product alternative referred to above requires the relevant CP to make what may be a substantial investment in infrastructure. This is particularly so for the SLU variant that applies to the NGA network, which requires a DSLAM in each street cabinet rather than in each exchange. There are around 1000 “tier 1” exchanges in the UK (that is, local exchanges), and around 100,000 street cabinets. A practical consequence of the shift to NGA has been that CPs have increasingly tended to opt to buy the bundled product, VULA, which is in practice replacing many LLU loops. As a result the number of lines operated under a LLU model has shrunk, and the reduction has not been materially offset by an increase in lines operated using the SLU model.

An additional potential consideration is Ofcom guidance against unbundling subloops in circumstances where a technology called “vectoring” is to be used. Vectoring adjusts transmissions in a subloop to alleviate interference between them, and does not work so effectively unless the entire subloop is controlled by one DSLAM. Vectoring is currently used on G.Fast lines and for a small proportion of VDSL2 lines.

The increased use of VULA has, in turn, reduced the extent to which CPs have been responsible for providing any DLM and have sourced DLM solutions from ASSIA. Sky stopped purchasing ASSIA’s DLM software in 2019. Between 2017 and 2020 ASSIA’s DLM related revenue from Sky and TalkTalk dropped by around 40%.

In August 2019 ASSIA became aware that Sky was providing ultrafast broadband services to its customers using G.Fast, in a way that ASSIA considered was likely infringing its patents. It approached Sky, initially referring to its use of vectoring technology, and inviting it to commence discussions about a licence of ASSIA’s patents in respect of that technology. Sky’s position, as developed in the correspondence, was that no licence could be discussed until ASSIA had settled the proper scope of its licence to BT. Following pre-action correspondence ASSIA commenced these proceedings against BT.

ASSIA’s position is that the terms of the Licence do not permit BT to provide services to third party CPs in the VULA configuration in circumstances where the CP provides any part of the necessary equipment, specifically the CPE (or more strictly, the modem incorporated in the one box version of the CPE), to the end user. It says that if this was wrong then, merely due to the rise of VULA, BT would be entitled to appropriate ASSIA’s entire DLM market in the UK. ASSIA accordingly seeks a declaration that, properly interpreted, the Licence did not grant any member of the BT group, including Openreach, the right to provide patent-protected DSL services on behalf of any CP which is not a member of the BT group to the CP’s end-user customers, where the CP directly or indirectly supplies any of the equipment necessary to provide the DSL services, such as any CPE. It relies in particular on clause 10.1 of the Licence, which is set out below.

In response, BT denies that the Licence is limited in the way claimed, and says that the effect of ASSIA’s construction would be that BT would have been in immediate breach of it, and the settlement would have failed to achieve its goal of achieving “patent peace”. BT seeks declarations to the effect that the Licence does not have a limitation that depends on the source of the CPE used on any particular line, that clause 10.1 of the Licence does not have the effect claimed by ASSIA, and further that the effect of the covenant not to sue in clause 8.1 of the Settlement Agreement is to prohibit ASSIA from bringing patent infringement proceedings against BT or (in respect of products purchased from BT) its wholesale customers.

It is convenient to consider at this point what ASSIA knew, at the time the Licence and the Settlement Agreement were entered into in 2015, about BT’s VULA product and its use, because that is part of the relevant factual context.

I accept BT’s case that ASSIA was aware at that time of the existence of VULA and that it was being provided by BT to CPs, including third party CPs. This conclusion is supported by evidence filed as part of the earlier dispute, including a witness statement filed in the Court of Appeal which listed a number of third party CPs to which NGA related products were being provided, and argued that any injunction could impact CPs’ services because they would not benefit from BT’s DLM.

In these proceedings Mr Trevor Linney, the Openreach research director, also provided evidence about the shift from a two box to one box solution. He explained in his first witness statement how BT has used the two box solution when introducing new technology, but that “wires only” (that is, one box) had been used in the 20CN and 21CN networks since around 2003. In 2015 the NGA network was “transitioning” to wires only, a process which Mr Linney said was complete by March 2016. He explained that what he meant by this was that thereafter all new NGA connections were provided on a wires only basis.

Mr Linney also explained that he believed that in 2015 ASSIA, like others participating in the UK industry, were well aware that CPE was generally provided by the CP, and further that the transition from two box to wires only was well reported in industry news and, so far as Mr Linney was aware, known to all industry participants. He further explained that Sky had publicly unveiled its own one box solution in December 2013, which was fully launched the following year, and that there was also discussion in an industry forum of which ASSIA was a member.

Professor Cioffi did not take issue with this evidence in his later witness statements.

I find that ASSIA was aware that CPE was generally provided by the CP, and was also aware that moves were being made towards using a one box solution, as had previously occurred with the 20CN and 21CN networks. ASSIA would also have had no basis to assume that BT was supplying VULA only to BT Consumer, rather than also making it available to third party CPs.

Clauses 3.1 and 3.2 of the Licence provide:

Clause 4.1 contains a licence of BT Patents on similar terms by BT to ASSIA in respect of ASSIA’s “Field of Use”. The provision about sub-licensing is however worded differently, as follows:

Clause 5 provides that the consideration for the agreement consisted of the “respective licenses and covenants not to sue set forth therein along with the monetary consideration set forth in the Settlement Agreement and Release”. Under clause 6.1, the duration of the agreement is linked to the expiry of the relevant patents.

Clause 1 sets out defined terms, including as follows:

It is worth noting that, although the licences in clause 3.1 and 4.1 are on similar terms, the scope of the patents and the respective “Fields of Use” covered is different. All of ASSIA’s patents are covered whereas BT’s patents are limited to ones related to DSM. (DSM, or Dynamic Spectrum Management, is a term that for present purposes can be taken to mean the same thing as DLM.) ASSIA’s Field of Use relates to its business of providing products or services employing DSM technology. BT’s Field of Use covers, broadly, systems used in or to control its networks, including DLM systems used in its NGA network. The differences reflect the respective natures of the parties’ businesses.

Clause 10 provides as follows:

The Settlement Agreement, which was expressed to come into force only when the Licence was fully executed (the “Effective Date”), provided in clause 3.1 for certain payments by BT to ASSIA in full and final settlement of the UK proceedings, less an allowance for a weekly sum that BT had been paying pursuant to an order of Birss J in lieu of an injunction. It also provided for additional annual payments in consideration of the Licence, the covenant not to sue in clause 8.1 (see below) and a commitment from ASSIA to provide certain public support and assistance to BT in respect of areas that included DLM and BT’s deployment of “G.Fast” in the UK. By clause 3.2.1 the annual payments (the last of which fell due on 30 July 2020) were stated to constitute “the remaining payment for all present and future use of any ASSIA patents”, as well as the commitment to support and assist and the covenant not to sue. Clause 3.7 provided that once the total agreed amount had been paid:

Clause 4 of the Settlement Agreement provided for the EPO oppositions to be withdrawn, and clauses 5 and 6 provided for the dismissal of the UK and US proceedings respectively.

Clauses 7 and 8.1 provide as follows:

The “Term” is defined as 20 years from, effectively, 27 July 2015. Clause 9 also expressly provides that the Settlement Agreement will continue for the Term.

Both the Licence and the Settlement Agreement are expressly governed by English law.

ASSIA’s position is that the provision by BT of DSL services (including DLM) as part of its bundled VULA product falls within the general scope of the licence granted by ASSIA to BT under clause 3.1 of the Licence, but where a third party CP is concerned it is excluded from its scope by the second sentence of clause 10.1, namely that the licence granted under clause 3.1 was:

ASSIA’s case is that the relevant third party is the CP. It maintains that: a) the relevant product or service is provided by BT to end users on behalf of the CP; and b) BT does so using materials provided by the CP, namely the CPE (which is typically provided by the CP).

ASSIA says that what BT is doing is delivering a service on behalf of the CPs. The CPs request BT to provide the service, and that service benefits both end users and CPs. An agency relationship is not required for that purpose.

Mr Segan relied on a number of cases that consider the meaning of “on behalf of” in various contexts, and in particular on a statement by Lord Sumption in Plevin v Paragon Personal Finance Ltd [2014] 1 WLR 4222 (SC) (“Plevin”), where the relevant issue was whether a broker’s failure to conduct a proper suitability assessment was conduct “on behalf of” the lender, so as to create an unfair relationship between the lender and the claimant borrower under consumer credit legislation. Lord Sumption said this at [30]:

Mr Segan relied both on Lord Sumption’s acceptance that the context may require the phrase “on behalf of” to be read more widely than agency, as meaning “in the place of”, “for the benefit of” or “in the interests of”, and on an argument that in any event the sense in which Lord Sumption was using the word “agency” was not limited to legal agency but extended to a “…‘thing’ … done or not done by someone whose acts or omissions engage the responsibility of” another person.

Mr Segan also submitted that the provisions of the Licence supported ASSIA’s case, in particular the fact that BT had no right to sub-license its rights to third party CPs under clause 3.2. BT’s interpretation effectively gave third parties the benefit of BT’s licence of ASSIA’s patents.

Further, the relevant contracts in this case were detailed and sophisticated, and the court must construe them principally by a textual analysis. However, if it were necessary to consider arguments based on business common sense or the consequences of competing constructions, then BT’s approach would allow it to trade away ASSIA’s entire UK market.

ASSIA accepts that its analysis only applies in a one box or wires only scenario, because where the two box solution is used the relevant box is the modem, which is provided by BT.

The principles applicable to contractual interpretation are not in dispute and are well established by the leading Supreme Court cases of Rainy Sky SA v Kookmin Bank [2011] UKSC 50; [2011] 1 WLR 2900 (“Rainy Sky”), Arnold v Britton and others [2015] UKSC 36; [2015] AC 1619 (“Arnold v Britton”) and Wood v Capita Insurance Services Ltd [2017] UKSC 24; [2017] AC 1173 (“Wood v Capita”).

Mr Segan referred me to the following summary in the judgment of Leggatt LJ in Minera Las Bambas SA & Anor v Glencore Queensland Ltd & Ors [2019] EWCA Civ 972 (“Minera”) at [20]:

There is a more detailed summary of the general principles established by Rainy Sky, Arnold v Britton and Wood v Capita in Carr LJ’s judgment in ABC Electrification Limited v Network Rail Infrastructure Limited [2020] EWCA Civ 1645 at [18] and [19]. I was not referred to that case in submissions, but I set it out because it is a helpful summary of those established principles:

Further points that are common ground are that the court may take into account a heading to a clause, unless another term provides otherwise: Lewison, The Interpretation of Contracts (7th ed.) at 5.108, and that evidence of pre-contractual negotiations may not be taken into account: Chartbook Ltd v Persimmon Homes Ltd [2009] 1 AC 1101 at [28]-[41]. It is also worth emphasising that the factual matrix that the court will take into account is limited to facts and circumstances which existed at the time the contract was made, and which were known or reasonably available to both parties rather than only one of them: Arnold v Britton at [21].

Mr Segan also referred me to Chitty on Contracts (34^th ed.) at 15-074 and 15-080 concerning the approach to construing related documents executed at the same time. Again, the basic principles were not in dispute. In particular, it was not disputed that the Licence and the Settlement Agreement should be interpreted together, reflecting the single transaction that they represent, and so far as possible consistently with each other. However, Mr Saunders rightly made the point that they remain separate documents, with distinct effects.

In my view a useful starting point in this case is to determine what the relevant product or service is that BT provides, being a product or service that might fall within the exception in clause 10.1 on which ASSIA relies.

What BT provides is almost certainly most accurately described as a service, but since it uses the term product and nothing material turns on that I will, where convenient, do so as well.

In the FTTC scenario relevant to this case, the product that BT provides is self-evidently the VULA product, that is a data connection which incorporates (1) the fibre connection between the exchange and the street cabinet; (2) the DSLAM; and (3) copper wires between the cabinet and the end user’s premises. It also incorporates a connection at the ethernet “layer”.

BT does not, for example, provide a DLM service. Instead, it provides access to a data connection between two points on the NGA network that includes an element of copper wiring, the quality of which may be enhanced by DLM as well as being affected by other matters, such as the length of the copper portion of the line.

BT also does not provide an internet access service. The provision of VULA would not, by itself, allow an end user to access the internet. It is clear from the evidence that considerably more is required to be done to allow that to occur. However, without the data connection provided by VULA it would be impossible for the end user to access the internet. It is therefore an essential element or feature. Further, the section of the line covered by VULA is in BT’s sole control.

These findings are consistent with the GEA Conditions (see [25] above). Under clause 3.1 of the GEA Conditions, BT agrees to provide the “Service” to the CP and to “grant the [CP] a non-exclusive non-transferable right to use the Service for the sole purpose of enabling the [CP] to provide the CP Service”, “CP Service” being defined as the service provided to the end user by the CP. The definition of “Service” clearly encompasses both the physical connection and an ethernet element, but does not extend to internet access. Paragraph 2.1 of Schedule 2B makes it clear that the “Service” of “Generic Ethernet Access over Fibre to the Cabinet (GEA-FTTC)” is to be provided by optical fibre cable between a handover point in the exchange and the “PCP” (that is, the street cabinet, being the “Primary Connection Point”), and by copper cable between the PCP and the end user’s premises (terminating at the BT network termination point there), with no handover to the CP at any intermediate point.

In order to succeed, ASSIA must demonstrate that the relevant product or service is provided by BT to the end user, since otherwise it could not be said that BT provides it “on behalf of” a third party (being the CP).

It is certainly the case that the end user benefits from VULA, because it is an essential part of their ability to access the internet. It provides the necessary connection between their home and the exchange, and indeed as far as the headend where the CP connects their backhaul (see [21] above). The quality of it, including the length of the copper portion of the line as well as the effectiveness of any DLM, will also have an impact on the speed and quality of internet access. The end user might even have agreed to pay more to the CP to achieve a faster connection. But as already discussed, VULA is only one part of what is required for an end user to access the internet.

What is provided to someone must logically correlate to what is received by them. An end user receives what he or she has contracted for, namely internet access. Although it might at least theoretically be divided into different services, that is not what the end user has bargained for or in fact received. If internet access is not available the end user will usually have little interest in the reason: they will just want it to be fixed, and will expect the CP to ensure that it is fixed, by whatever means are necessary. (In many cases an end user may have contracted for additional services, such as email services or subscription TV, but those services are not relevant to the issue that I need to decide.)

Various possible analogies were discussed at the trial, the one that Mr Segan relied on in reply being Royal Mail sub-contracting the delivery of mail from the local sorting office, including posting it through letter boxes. I did not find any of these analogies to be very helpful, but would comment in relation to the one just mentioned that it is far from obvious that the service provided in that case is anything other than an end-to-end delivery service, which cannot sensibly be sliced up between different sections as far as the customer is concerned. Further, on the facts of this case the CP’s role is not limited to ensuring that messages are conveyed along a line, but ensuring that effective internet access is achieved.

In contrast, it is clear that VULA is a product provided to, and received by, the CP. The CP must have the benefit of it to be able to provide the service that it has contracted to provide to the end user. If it were required to enforce its contract with BT it would be doing so for its own benefit, to allow it to perform its own obligations to end users. It would not simply be seeking to compel BT to provide a product or service to end users in its stead.

I do not accept Mr Segan’s submissions that various provisions he referred me to in the GEA Conditions, such as an obligation owed by BT to repair faults or for the CP to take steps to ensure that end users prepare sites for connection work, support ASSIA’s case. If anything, they tend to emphasise that BT provides a service to the CP.

It is worth noting that the end user would typically have no knowledge of or interest in the arrangements between the CP and BT. They may appreciate that they benefit from the faster speeds available with FTTC, but they will usually be both ignorant of and indifferent about whether that is achieved using VULA or SLU, the choice between them being one for the CP to make. With SLU, the copper wires leased to the CP are just as essential for achieving internet access, and BT will no doubt be responsible for their maintenance (and indeed installation where required). But it would be unrealistic to say that the copper line is provided to the end user on behalf of the CP. It may be commissioned from BT by a CP, if not previously installed, but once installed the line would be available to any CP with which the end user chose to contract.

Mr Segan relied on the fact that, as the GEA Conditions and Mr Linney’s evidence make clear, BT has full control of the portion of the connection covered by VULA, there being no handover to the CP at any intermediate point. However, this does not affect the fact that the service contracted for and received by the end user is the single service of internet access.

In summary, in my view BT provides the VULA product to the CP, and not to the end user.

Not surprisingly, the cases relied on by ASSIA demonstrate that the words “on behalf of” may encompass a range of meanings which may differ according to the context. Mr Segan submitted that, even adopting the narrowest construction, BT’s provision of VULA fell within those words.

I am prepared to accept that, if I was wrong and the correct factual analysis was that BT did provide a product or service to the end user, then it would not be precluded from doing so “on behalf of” the CP by the absence of what in law would amount to an agency relationship. I agree that, in the particular contractual context of this case, those words do not necessarily require such a legal relationship to exist. The comments made by Lord Sumption in Plevin at [30] about acts or omissions that “engage … responsibility” are apposite. Were it otherwise, the provisions of clause 10 might be relatively readily avoided, such that it would not achieve its evident purpose (as to which see further below). Using Lord Sumption’s terminology, BT’s actions or failures would “engage the responsibility” of the CP, because the CP is responsible to the end user for the existence and quality of the service that it has contracted to provide.

As already indicated, ASSIA’s case is that where the CP provides the CPE in the “one box” solution (that is, including the modem) then BT provides a service to the end user using that CPE. I can see that, if BT were regarded as providing the data connection that comprises VULA to end users, then a necessary part of that is the modem at the end user’s end. On a literal interpretation of the words this requirement could therefore be satisfied.

Whether this is the correct interpretation of the words in their contractual and wider context is considered further below, but one relevant point to make at this stage is this. Whether or not CPE is provided by a particular CP or by someone else has no particular relevance to what BT provides. The VULA product just requires a functioning modem, operating to the relevant technical standard. It would at least be technically possible for the end user to source their modem from someone other than their CP, or to re-use one previously sourced from elsewhere. The CP may well have no knowledge of that, and neither would BT unless it supplied the modem itself. Further, although there are commercial reasons for one box solutions to be preferred by CPs (and no doubt by most end users), that preference has nothing to do with the VULA product that BT supplies. Therefore, if ASSIA’s construction were correct then whether BT was in breach of the Licence would depend on what, from the perspective of its provision of the VULA product, is the happenstance of a choice being made to use a CP-sourced modem. Further, that would be the case even though that modem plays no active part in the DLM processes that were the focus of the dispute between the parties and that are captured by the “Field of Use” definitions in the Licence.

Mr Saunders submitted that the final reference to “provided” in the second sentence of clause 10.1 (that is, “using materials provided by or on behalf of the third party”) refers to materials provided to the relevant party, that is the licensee, rather than to the recipient of the product or service. I can see that this is a possible construction, and one that would fit with certain kinds of transaction that may well be regarded as patent laundering, but it is not straightforwardly correct. The two earlier references to “provided” in the same sentence clearly refer to a provision by rather than to the relevant party, and it is by no means obvious that something different was intended with the third reference.

One of the issues at trial was whether, or the extent to which, I should have regard to witness evidence from Mr Peter Ratcliffe, a senior member of BT’s in-house legal team, about the concept of “patent laundering”. Among other things, his first witness statement dated 16 December 2020 sought to explain the legal background to the development of patent laundering clauses in the United States, what their scope was intended to be and why clause 10 was not engaged in this case.

In oral submissions Mr Segan submitted that BT could not deploy this evidence because evidence of foreign law needed to be proved by way of expert evidence. Mr Saunders responded that this point was raised far too late, but in any event that the courts’ approach to proving points of foreign law was no longer as strict as it once was, referring to Lord Leggatt’s observation in Brownlie v FS Cairo (Nile Plaza) LLC [2021] UKSC 45 at [148] that:

I do not need to rely on the evidence that Mr Segan objected to in order to reach my conclusion. What is clear from the factual evidence, and is obvious from the heading of clause 10, is that the parties understood that a concept of patent laundering existed and at least in broad terms what it typically involved, and that they intended to address it.

Clause 10 itself contains clear clues about the sorts of transactions that the parties must be taken to have had in mind. The first point to note is the heading, which starts with the word “Clarification”. That is an indication that the parties intended to make clear that there were certain transactions that the licences were not intended to cover, in case it was argued that they were within their scope. It does not carry the flavour of a clause that is intended to make a radical alteration to the scope of the licences, but rather to spell out something that might otherwise leave scope for argument.

The first sentence of clause 10.1 refers in terms to patent laundering, and describes what must on any basis be a classic example of a transaction answering to that description, namely the holder of a patent licence undertaking manufacturing activities for a third party, in a way that they no doubt hope will benefit from the licence. Clause 10.2 describes another “laundering” example, where products are bought from and resold to a third party, again no doubt with the aim of conferring the benefit of the licence on those products.

The second sentence of clause 10.1 commences with the word “Similarly”. That provides an indication that that sentence is intended to address a similar vice to that described in the first sentence, or at least to address transactions that bear some relationship to what is covered by the first sentence. A paradigm example, applied to the facts of this case, might be one suggested in BT’s skeleton argument, namely a situation where a third party that does not have a licence from ASSIA provides a DLM solution covered by ASSIA’s patents for a CP’s LLU or SLU lines, but asks BT to operate the DLM on its behalf in an attempt to benefit from BT’s licence. BT would then be providing DLM to the CP on behalf of the unlicensed third party, using materials provided by it, in an attempt to “launder” the unlicensed system using the licence held by BT.

Another example, which perhaps reflects why BT originally asked for the clause, would be ASSIA attempting to extend the benefit of the licence it holds from BT to a third party’s DLM system, by providing DLM products using BT patented technology on behalf of that third party.

BT relied on the third sentence of clause 10.1 in two ways. First, Mr Saunders relied on the references to Fields of Use and the “normal course of business” as indicating that the clause was getting at something out of the ordinary: it was not supposed to infringe on normal business operations within the parties’ respective Fields of Use, which were themselves defined by reference to their different business operations. I agree that this is of some relevance to the construction of clause 10 as a whole, especially given the introductory words “For the avoidance of doubt”.

Secondly, Mr Saunders submitted that even if what BT was doing was caught by the second sentence of clause 10.1, that was cured by the third sentence. This was on the basis that: a) BT continually modifies its networks, and further customises VULA by providing it to the particular end user in question; and b) BT provides the VULA product to CPs under BT’s branding and trademarks. Whilst I see these points I am not persuaded that they assist BT if the VULA product could (contrary to my view) be said to be provided to end users. This is because, as far as they are concerned, they do not receive a BT-branded product, and indeed they may not even know that BT has any involvement. Further, whilst it is no doubt the case that BT does modify or improve its products on an ongoing basis, a more natural interpretation is that the reference to “modified, supplemented or customised” is getting at some form of alteration or addition made for the benefit of the customer in question, whereas the VULA product is provided in the same manner to all CPs.

Mr Segan submitted that if it was correct that BT did not provide VULA to end users, then an example used by Mr Ratcliffe in his evidence of what clause 10 was intended to capture would not fall within it. This was an example of BT using a DLM solution provided by an unlicensed third party CP to control that third party’s DSLAMs (that is, in a LLU or SLU scenario) for the benefit of end users.

I do not need to provide a definitive answer to this hypothetical scenario, which was not ultimately relied on by Mr Saunders. However, one potential response is that it would not fall within clause 3.1, properly construed, because it would not fall within the BT “Field of Use”, and would also not be regarded as part of BT’s own “networks, products or services” for the purposes of the covenant not to sue in clause 8.1 of the Settlement Agreement (as to which see further below).

As already noted, the terms of the licence granted in favour of BT under clause 3.1 of the Licence allow it to use ASSIA patented technology in a Field of Use that explicitly includes DLM systems operated in its NGA network. No limitation is placed on this. It is true that under clause 3.2 BT is not permitted to grant sub-licences, but ASSIA rightly does not seek to argue that the operation of the RAMBo on BT’s NGA network amounts to sub-licensing the technology to any CP who, or whose customers, benefits from that operation. If ASSIA’s argument about clause 10.1 were correct there would be a very material restriction to the scope of the licence which is not even hinted at in clause 3.

It is also worth noting that clause 4.2, which deals with sub-licensing by ASSIA, includes a specific proviso to ensure that the licence remains attached to products and services “properly” sold or licensed “within the normal stream of commerce”. It is apparent from this that, as with the third sentence of clause 10.1, the intention was to allow the parties to carry on what was understood to be their respective normal business operations. ASSIA’s normal business included supplying DLM software to third parties, and it was obviously intended that that should be covered by the licence. BT’s normal business includes the operation of its NGA network.

As already indicated, there was no dispute that the Licence and the Settlement Agreement should be construed as a single “package” of documents.

The release in clause 7 of the Settlement Agreement is drafted in broad terms. Clause 7 is supportive of BT’s case, discussed further below, that the parties intended to end the dispute between them rather than to perpetuate it through a restriction in the Licence which would have immediate implications for BT’s business. In particular, the agreement is stated to be in “full and final settlement of all disputes … as at the Effective Date”, and the terms of the release of rights are broadly drafted, covering claims etc. “arising out of or connected with” the litigation.

Mr Segan relied on the presumption that neither party intends to abandon any remedies for a breach which arise by operation of law, and that clear express words must be used to rebut that presumption: Gilbert-Ash (Northern) Ltd v Modern Engineering (Bristol) Ltd [1974] AC 689 (HL) at 717, per Lord Diplock. However, in Transocean Drilling UK Ltd v Providence Resources plc [2016] 2 All ER 606 at [21], Moore-Bick LJ made clear that Lord Diplock’s observation must be read in the context of that case, and that any presumption must give way to the language of the contract. In this case it is clear that the parties each contemplated giving up material rights.

The covenant not to sue in clause 8.1 is also drafted broadly. Indeed, at least on its literal interpretation it appears to prevent action for any patent infringement at all against the parties or their respective customers, irrespective of whether it has any connection to the dispute that was the subject matter of the settlement and also irrespective of the terms of the Licence. Mr Segan submitted that it could not be intended to cut across the terms of the Licence, such that properly construed it did not prevent ASSIA from asserting that BT was in breach of it by virtue of clause 10.1. Mr Saunders responded that on ASSIA’s interpretation of clause 10.1 the provisions conflicted, whereas on BT’s they could be read consistently. This was because clause 8.1 only applied in respect of the “networks, products or services” of the other party, rather than products or services “laundered” for a third party.

Mr Segan’s submission that the parties cannot have intended that the covenant not to sue would prevent enforcement of the terms of the Licence, and so should not be read inconsistently with it, has some force. However, that does not mean that the Licence should be interpreted divorced from the Settlement Agreement and the wider context. In particular, it does not dictate that one potential interpretation of clause 10.1 of the Licence should prevail over any other.

Overall, considerations of business common sense and the relevant factual context favour BT’s case.

First, it is significant that ASSIA’s case depends entirely on whether it is the CP that provides the end user’s modem. If BT provides it (as it does in the “two box” case) or it is sourced from elsewhere then ASSIA accepts that BT’s provision of VULA falls within the terms of the Licence.

There is no rational relationship between the source of the modem and whether the Licence applies. It is the case that one box solutions are preferred by CPs, because they are neater for the end user and facilitate brand recognition, but those factors have nothing to do with VULA. They would be present whatever type of DSL product was used.

Further, the fact that the existence of a breach would be dependent on the end user using CPE provided by the CP means that whether BT was in breach would ultimately depend on something outside the control, or even knowledge, of BT or the CP, namely the end user’s own choice of equipment. That lacks logic.

In contrast, an interpretation of the second sentence of clause 10.1 that does accord with business common sense is one where the use of “materials provided by or on behalf of a third party” has some real relevance to the activity complained of. The closing words “and then to the extent that such products or services are provided using materials provided by or on behalf of the third party” are clearly intended to operate as a limit on the scope of the provision, in the sense that there will be no breach if or to the extent that such materials are not used. The parties cannot readily be taken to have intended to impose a limitation which is wholly arbitrary and appears to have no commercial purpose. The closing words strongly indicate that the vice intended to be caught is one that is linked to the third party’s materials being used. That approach would also be consistent with the first sentence of clause 10.1, and with clause 10.2: both are clearly getting at products manufactured for or bought from a third party.

Secondly, it is important to bear in mind the nature and content of the dispute that the Licence and Settlement Agreement was intended to bring to an end. What was in issue was the respective parties’ DLM technology, and specifically in BT’s case the RAMBo used on its NGA network, which had been found to infringe ASSIA’s patents. Unlike ASSIA, BT did not (and does not) operate a business of supplying DLM solutions to CPs as such. BT uses DLM technology only to control its own DSLAMs, and in the case of the NGA network it does so, and did so in 2015, as part of the provision of its VULA product.

So the significance of the licence granted under clause 3.1 in BT’s favour was that it allowed it to continue to provide VULA, with the benefit of DLM, free from infringement claims by ASSIA. The absence of any ability to sub-license, in contrast to the more nuanced provision in clause 4.2, simply reflects the parties’ different business models.

VULA has existed since 2009. I have accepted BT’s case that when it agreed the settlement ASSIA was aware of the existence of VULA and that it was being provided to CPs, including third party CPs. Further, I have found that ASSIA was aware that CPE was generally provided by the CP, and was also aware of the transition to using a one box solution. These facts are relevant factual context.

Thirdly, it was also not in dispute that the aim of the Settlement Agreement and the Licence was to achieve “patent peace”. I consider that this is apparent from the contractual documents, but in any event evidence of the genesis and aim of a transaction is admissible: Merthyr (South Wales) Ltd v Merthyr Tydfil CBC [2019] JPL 989 (“Merthyr”) at [43], [52]-[54] per Leggatt LJ, referring to the classic statement of the principle by Lord Wilberforce in Prenn v Simmonds [1971] 1 WLR 1381 at 1385 and subsequent cases. I note that in Merthyr at [50] Leggatt LJ disapproved an observation by Sales J in Investec Bank (Channel Islands) Ltd v Retail Group Plc [2009] EWHC 476 (Ch) (“Investec”) at [76], insofar as it suggested that evidence of negotiations could be admitted to show the genesis and aim of the inclusion of a particular provision in a contract, as opposed to the genesis and aim of the transaction as a whole. BT had relied on Investec in its skeleton argument but Mr Saunders rightly accepted that it needed to be read in the light of Merthyr.

If ASSIA’s construction were correct then the settlement that the parties purportedly agreed would, in truth, not have ended the dispute and achieved “patent peace”. On ASSIA’s case BT would have been in immediate breach of the terms of the Licence in continuing to provide the VULA product to third party CPs. Such a construction would require very clear words, and indeed I would have expected a clear carve out to be contained, or at least signposted, in clause 3, rather than in a single sentence contained in a “clarificatory” clause.

The clear commercial object of both parties was to achieve commercial certainty. The Licence was part of that, because it covered acts that the parties were otherwise alleging were infringing.

I do not accept Mr Segan’s suggestion that BT’s confidence that it had modified its NGA RAMBo to avoid any infringement supports ASSIA’s construction of clause 10.1 (on the basis that if BT was correct it would not require a Licence). It certainly does not follow from the fact that it was expressing confidence and seeking summary judgment that it would see no benefit in settling the dispute, and doing so on a basis that made clear that any arguable infringement was expressly authorised.

ASSIA’s complaint that BT’s construction would enable it to appropriate ASSIA’s entire DLM market in the UK is really an invocation to apply commercial common sense retrospectively, rather than at the date the contract was made. Since the date of the contract the market for VULA has grown substantially as usage of FTTC has increased, with SLU being less favoured by CPs. However, the fact that this may not have been anticipated by ASSIA is not relevant. It simply means that it has concluded what, in retrospect, it may regard as a bad bargain.

I have concluded that ASSIA’s claim should be dismissed. As proposed at the trial, I will address the declarations sought by BT following submissions made consequentially on this judgment.