Industry sign banking rhode island document computer
my name is Gordon Hughes I'm the global Sales Director for financial services that alliances at Intel and I'm sort of going to do brief injection and hand over to my friends c9 who's going to talk more about what market logic are doing in financial services so just by way of a brief introduction you may be wondering and I'd get this quite office but why is Intel here Intel does processes don't they well yeah we do processors but we do an awful lot more than that you know we have we have processing technology including of it you may be aware that we recently bought alteracao we have FPGAs we have CPUs we have many cored CPUs we have storage technology including solid state drives non-volatile memory we have networking fabric we also have a lot of software as well mainly software tools one thing that you may not know is that if the software business in Intel was taken out as a separate business then it would be the third largest software vendor in the world so we have a real broad spectrum portfolio technology so because of that one of the things that we need to do when we engage with customers is understand the business issues because we're not just selling point products we need to understand what the business issues are and then we need to work with our partners to develop and define business solutions that absorb all our technology and provide you know a complete business solution to the customer so with that in mind and now turning to the financial services industry that's the financial services industry has always been a big focus industry over the years anyway because it's probably you know the the biggest industry when it comes to high performance compute everything from the investment banking side through you know the low latency trading high frequency trading risk calculations and there's just so much happening in that industry that can leverage Intel technology if we step back and look at the pressures on that industry at the moment there are two major areas of pressure one this comes as no surprise at all and we all know it that the financial services industry here pressure from the regulators that was clear from the conversations in the KPMG session yesterday and we're all acutely aware of that the numerous regulators and regulations that are hitting that industry the other thing that is putting a lot of pressure on the industry is competition there are dissing to disinter mediators there are challenge of banks emerging now whose has anybody heard of PSD to the revised payment services directive that's been pacified the European Parliament I won't go into the detail of that but that is going to cause huge disruption in the financial services industry because it's going to open up that industry to all sorts of third parties who are going to directly compete with the banks so there's those two huge pressures hitting the industry so if we look at the trends in the industry or the themes in the industry we as Intel have identified three major themes or activities in the industry one is around digitization so that is taking non digitized processes and digitizing those and we heard about that in Michael Henry session and the KPMG session yesterday and what they're doing in terms of digital labor looking at kyc there's a whole area of transformation in the data center so thinking about you know the sort of migrating off older legacy systems aggregating and consolidating data silos we heard about that in the keynotes yesterday from David at marklogic and the third area is around innovation and this is a whole area of data management and analytics there's a lot of area a lot of activity in the FinTech space with new types and analytical processes you'll hear conversations about neural networks deep learning machine learning and all those sort of things so that all those things are hitting in the industry and it's interesting because all those three areas that are areas that are being addressed by MarkLogic if we look at the innovation area in terms of data management and analytics and we're going to focus on one particular area of that around regulatory reporting in this session you know all that all the three areas I've just talked about have been covered this week in the session we're hearing from marklogic so with that in mind I'll pass over to see now now to focus in on some of the work that that Intel and MarkLogic have been doing to jointly address the issues around regular reporting in the financial services industry see No thank you just remind you of the environment we were all working in most of the information that flows in the financial system is increasingly being subject to regulations some of them didn't exist as early as six seven years ago and some of them are being enforced far more diligently and across borders it's creating a requirement to not only move the data much quicker but retain a lot more data in a manner that is accessible to the bank's the auditors and regulators and the volume of the data is and the variety of data is increasing and it will not end any time soon as the new regulations like the one Gordon mentioned is going to mount and that produces a lot of stress in the data management infrastructure the imperative is real and it is being especially in Europe enforced with a certain diligence and consistency that we haven't seen before in the 90s in the in the earlier times Basel 3 it probably going to be succeeded by Basel 4 not to negate what was required before but also add to it and the roles and responsibilities and the way things are done the regulatory cycle is being managed by the banks is also being regulated it's not only what you do when you do it but how you do it has become on the regulatory purview and the answer is essentially comes down to what you use technology and how you automate and how you in accuracy security and accessibility and preferably in one homogeneous infrastructure the data is not homogeneous to any degree that would make it easy and therefore there are intelligent solutions that need to be brought to bear to understand the variety of the data in the intrinsic value of different types of different formats of data but still be able to go through the aggregation path and produce the reports on time so when we look at the audience we address ourselves to with Intel in mark logic we have two CEO who's really varied about growth top line bottom line growth CEO execution CIO enablement of the corporate functions and the chief financial officer and the Treasury of sorry about the risk exposure of the balance sheet risk of the the capital and the operational risks including reputational risk let me look at the return on equity in sector wise banking doesn't rank at the top it's still under financial stress but it is not as bad as it used to be four or five years ago it is not in the 2% 3% or even negative anymore it's reasonably you know recent history wise relatively healthy 10% and as that you know balance sheet health improves hopefully in the years to come there will be additional investment available to streamline operations and become more compliant become more automated and provide more secure and highly performant platforms for the movement of data and management the risks so regulations essentially impose an imperative to invest in the transparency of the information that flows make that information visible internally and to the regulators but running the business is not any easier there's been tremendous consolidation in the industry banks are bigger they have absorbed other organizations and through mergers and acquisitions have inherited even more diverse and heterogeneous IT infrastructures and they have to manage that complexity and one of the ways to do that is disintermediate the fragmentation of information eliminate silos and that's where we play very strongly and could have claim to both competence and scale in terms of eliminating silos at least reducing them innovation is ever-present the larger consolidated bigger enterprises now competing for the customer far more diligently and product innovation innovation in services is the way to acquire new customers they have to acquire one another's customers incent cross-selling and make the customer a more complete more involved customer for their offerings so being able to deploy and to market in time and ahead of the competition is the third imperative and the customers not only to be able to cross sell to the customers not only to expand the wallet share of the customer but also to be able to account for what the customers do partly as part of the regulations but partly as part of understanding the customer has also become an imperative and that's the fourth one be able to profile the customer properly so you can understand and service the customer if need be be able to audit and account for customer behavior and customer transactions to the regulator's and litigators if need be so given that we have a solution architecture with mark logic to facilitate both internal governance and regulatory reporting and we rely definitely on the underlying technologies and the most important one is the technologies we rely on from Intel so I will let Gordon go into the Intel technology as the foundational technology that we all rely on okay so yeah so this this this is really what we've been doing with mark logic and other partners in different industries we're looking at solution areas and then working with folks like mark logic to define solute solution plays or solution architectures that address you know sort of key hotspots in the industry and the big one that we identified with Mark logic is around regulatory reporting so if we sorry so if we go into the next slide this is a high-level view you know this is just representative as sort of issues that the banks are facing so so you know if we look at the sources of data here there's data coming from the trading systems there's event stream processing market data reference data streams entering the trade repository which is typically a tiered storage you can store your trades there your counterparties your reference tables and legal documentation the banks have huge high performance grids doing risk calculations on these these can be tens upon tens of thousands of processing cores acting on that grid and in here there's the need to to look at this data and now correlate this data for the purposes of regulatory reporting so what the banks now have to do they have to take their legal documentation and they have to report on the trades that have been executed by the counterparties against the legal contract between those counterparties and this is actually quite complex reporting requirements which is very very well serviced by in mark logics no SQL database there are analytical components that are required SAS analytics is just one example of the components there and then there's the need to build a regulatory store when I first started talking about this solution both to mark logic and also to the bank's what what amazed me was that the most of the banks today do not have a regulatory store they produce reports from their trade repository send the report to the regulators but what they haven't been doing is actually capturing those reports and storing them in a regulatory store so they can go back to them later on and again this is one of the key things that mark logic can facilitate XBRL extensible business reporting language is becoming the de facto reporting language the regulators and again this is something that is very well facilitated by the capabilities in mark logics no SQL database there is a need to store that and then maybe revisit that and report back to the regulators and that would be submitted to the different regulators in the different countries and it's not one regulator per country right because you have you know many different types of regulation per per country and all that needs to be reported out at different time intervals to the regulators via you know daily weekly monthly or annually the other thing that causes real pain in the banks is stress testing or had ad hoc reporting because the regulators can go back to the banks at any point in time and said I want you to rerun your reports right and I want you to change the parameters against the reports assume a different rate of inflation assume this assume that for those of you who live in the UK you've made seen some of this happening about a year or so ago when some of the made three of the major banks I think was RBS Lloyd's and the coop Bank were the the regulators demanded a stress test their systems and I think RBS just scraped through by the skin of the teeth but correct me on that I think co-op Bank actually failed on that that's what was happening that regulators were going back to the banks and asking them to rerun their reports and stress test their systems with different set of parameters and that's a particularly painful process that the banks have to go - you know it's very very expensive and painful so this is really why we put worked with MarkLogic to put this architecture together to address those particular sorts of issues that the banks the banks are facing you know what's not shown on here is the possible ingestion of things like data from social media sites the volume of data that needs to be ingested here is only going to grow there are going to be more and more sources that the banks need to ingest I think just the legal documentation alone can absorb that to a petabytes of storage this is yeah this is this is huge amounts of storage again which MarkLogic is is very well-suited to to cope with as you have done so both in finances but other industries as well so so one of the things that we do when we work with MarkLogic is actually make sure that they've taken advantage of the latest into Intel technology so there's a bit of a plug for our latest products our latest version of the xeon processor family we work very closely with the market logic development teams and make sure that their products are optimized to run on Intel technology and here we have quotes here from kayo Milani who's the director of product management of our logic citing the way mark logic has been working with Intel both to use our solid-state drive technology and the latest Intel Xeon processors a lot more information about this some graphs here about how our latency on processes are actually increasing the performance of queries and the market logic database I'm not going to go into this in technical detail if you want more information please talk to myself or seen on offline we can give you a wealth more information but I'm not going to turn this into a product pitch anyway on that note I'll hand back to seenin ok thank you very much so based on what you listen to about the generations of Intel processors and the the storage devices that we use we have the high light capabilities that we have to deal with the entire regulatory and one of them is the scale-out commodity hardware other than the deploying units are thinking right now it's eight core units of server that gives you some vertical scalability but as you would appreciate we need the horizontal scale out and that's where Intel helps us leverage out that capability we can use commodity hardware and be able to extend the computing capacity as the volume of data and the processes on that data become more and more complex and the volume continuously increase the tiered storage gives us the capability to actually establish some policies about age data and since now we are dealing with the requirement to have data deep archival data available on demand to auditors and regulators and that demand is not going to go anywhere we can't wait a couple of days to harvest the data historical data and present it for investigation or litigation if need be it has to be accessible and online that's an expensive proposition if you keep everything real-time accessible but actually you can deploy aged data to lower priced hardware and storage devices and keep the near real turn data on a faster access media then you can actually establish some very intelligent aging policies for very large amounts of data and still meet the requirements semantics gives us a an entirely different
way of accessing data and make inferences based on relationships our ability to transform data into RDFS and apply semantics and ontology to it makes us a lot easier it makes it a lot easier for the users to establish the relationships that are not all that obvious when you look at the raw data via transaction compliance so if you're collecting data from a transactional flow we will maintain the transactional integrity as the unit of transmission of data that is coming in and be able to reconstruct that transaction demand we look at the data for investigative purposes we do not lose the that unit of work concept of a transaction let me capture it from a transactional flow by temporal capability is it's one of the very strong differentiators of mark logic right now is an or sequel platform it allows us to ask the question who knew what when we can establish the data events on two timelines and be able to go back and reconstruct how that particular information was used at a point in given time in what context and we never lose that temporal relationship between different timelines geospatial has been around for a while and MarkLogic is a capability came out of law enforcement and intelligence requirements intelligence community's requirements now we can correlate physical location with data events with movement of data and some analytics on that data and be able to give you a time and physical space based influence on what is happening when and where real-time alerting is our ability to actually establish and automate an alerting mechanism based on data events in relationships between data events so that we can notify the user that there's some anomaly if you would or some reinforcements of some behavior that is under observation and be able to lead them to more thorough analysis of what has been happening and we can do that by establishing alerting rules at ingestion time and government great security we've been certified to be a secure platform by the by the federal government we have encryption policies and algorithms applied to the data at ingestion time and when the information is committed to the storage device and we run through the certification process and I think again I think we are the only no sequel platform that has received that degree of certification if not mistaken if I'm mistaken Kent will correct me so most of the data that regulatory process requires right now it seems many is going through many hops and that is wasted capital and time and valuable time is being lost to validate the data in transit and then it arrives and we are trying to reduce that cost and the time and we would like to capture some of this 36 billion dollars spent on innovation and new applications rather than catching up with the data itself so relational databases are good if we're dealing with pure or ltp transactions and we need that transaction flow come in but that doesn't mean we have to live with the encumbrances in the limitations of many steps of ETL and the risk of actually corrupting data in due process or D synchronizing data flow from related data flows and try to establish that temporal relationship later on the database solution that we offer to integrate faster and provide access searchable queryable database leverages essentially two strengths of the Intel technology the processing speed and access to memory we rely on availability of ram space addressable space within our server processes in order to not only accelerate the data processing but also be able to keep data in full view as we were manipulating it and validating it and everything I described to you up until this point is taking advantage of the increased clock speed of the processors the multi-core architecture of the processor so that we can fragment the workload within a processor across cores and then be able to use the the addressable memory that is available to the processor in each core in some very intelligent way so that they can actually keep both the data flow moving and the data interesting processes we have in sync so our solution is address the the data capture problem regardless of what format in shape the data is in we will capture it it will ingest it we can get the data and the relationships between data out through a search engine in the query facilities and the semantics that we have and we can keep it secure wholesome a holistic image of the data and control the the access by different user groups with access protocols and policies by doing that we will be able to deploy in on-premise or in the cloud same and capabilities and be able to use AWS in Azura's our deployment partners we can interface with other technologies like Hadoop we can interface with the business intelligence platforms for visualization and analysis so those are our credentials for a regulatory reporting cycle as well as internal governance and auditing we are giving a complete and coherent view of the data and be able to maintain the temporal relationships between data events across the both the transactional streams of data and harvested data from social media and other streams and other sources so that that concludes my presentation I'll entertain questions I'm sure Gordon will be taking questions and comments right well it would be that's right well I mean it is a qualitative sort of assessment really I mean the conversation how with the banks is that you know have the conversation around what is the pain of stress testing or doing the ad hoc reporting because they're going to get more and more of that coming down from the regulator's yeah across yeah that's right the other thing so the other thing I didn't discuss on that particular slide was the fact that just just to keep it simple there's almost a mirror image of part of that that would be implemented by the regulator's themselves in their own data centers because the regulator's themselves are requires of Technology so you know they're going to receive those those reports they're going to have to do their own analysis on those reports and then they as I understand that they issue the requirements for for ad hoc report of stress testing back to the bank in an XBRL format so there's a vicious circle going on there you know we know that most of the regulator's as I say are are acquiring technology so you know from a vendor point there's an opportunity to engage with the regulator's there I mean the other thing about the regulator's and technology is the regulations themselves can actually somewhat vague when it comes to implementing technology solutions that the regulations themselves are can be very very fuzzy I mean one of the things about I mentioned PSD to one of the things the banks are struggling with at the moment what on earth does this stuff mean in terms of technology if the regulation is saying I need to open up my systems to allow a third-party payment providers get access to my the the customer data that I hold for my customers how do I do that now a lot of people talk about that in terms of api's know where in PSD to is the term api ever mentioned all right so there's a lot of second-guessing by the banks that goes on when it comes to interpreting what what those regulations mean from our point of view right the important thing is flexibility you could have an agile and flexible solution right you're not going to achieve that through a traditional relational database right what you are going to do is be able to achieve that through through using marklogic no SQL in fact some of the regulations are even almost if you take cup I can't give specific examples but I know talking to people some of the regulations are almost contradictory one regulation will say something about you know knowledge of customers and another regulation would say that something that seems to be diametrically opposed to that regulation over there it's yeah it's the the I mean if you but we should presume some technological depth on part of the regulator's when they make these regulations and explain it in paragraphs and words but if you go back to one of the earlier quotes from McKenzie research where they almost imply data lake or data lakes as the medium where you'd go to get your first order access to information but the proviso of that is whatever it is data warehouse or data Lake it has to maintain the progression of data and the modifications and transformations of data in time so that if you want to get the chronology of something you're investigating it has to be unraveled with the time dimension and then you'll be able to say who did what when and if it's the wrong thing or inadvertently done because the data was corrupt or wrong or incomplete if you can retain if not right if the requirement implicit requirement is that degree of capability in granularity it is not just piling of data in one place that you can get all the data accessible but how did that data became that data from the origin to the point of use that that's what they're interested in I can first of all fibro is an initiative that may succeed or be adopted or not but beside fiber the use of semantics in ontology there are two ways to think about it one is define a vocabulary in taxonomy about things that people use like a bond or stock and then accumulate a language to describe what bonds do for a person or an institutional investor in what an equity does as part of an allocation model in a portfolio or from a regulators point of view you look at processes and the banks CFO or treasurer is going to look at say liquidity management as a process so how do i aughtta mate all the decisions that I make around liquidity management using ontology then you're a very different approach to building ontology I'm not going to build an ontology to describe financial instruments or accounting concepts I'm going to describe an ontology that feeds into a well-defined process I may have a different ontology for liquidity risk management and clayman in the liquidity management or collateral management that is the difference in approach if I were a regulator I'm going to see what is ontology going to do for me I'm gonna say it will make me understand the process better automate the process and be able to make that process screw Tabo you know that I can go and analyze at a very granular level technologists impulse or and academicians impossible to start with the words we use and then enrich them through ontology ascribe some meaning to it then I say stock I don't mean the stock in this store but I mean shares or the company and then dividend and all of those things you can build an ontology around it so that's the technologists impulse to build up ontology the regulator's would really would look at a process yeah and and start there and they say is it controllable is it governing oh is it suitable is it automated and is it going to keep the data all of the data in time the the other what was the first question was about the question you asked before you mentioned fiber yeah no I mean there is no government entity that I would call a regulator in the five oil activities today very few banks even there about only three of them that appear to be sponsors a lot of it is driven by technology companies or consultants or academics so I don't know if regulators actually are aware that there are ontology building societies emerging all over the place yes some work going on I mean there's you see if things like heat and if you were the EDM council yeah so that's a nonprofit organization and that this is something they're trying to address they're trying to bring the banks together largely the chief data officer Xin the banks and also have some interaction with the regulator's to define common ontology so they're working on that it wasn't not necessary specifically to address the issues around negative reporting but but it is a work stream and they're looking at all sorts of different things there I mean it's it's a worthy venture we'll have to see how much yeah and if you look at the board of directors for the EDM counts is like a who's who of the chief data officer Xin the major banks all right how much traction they're getting with the regulator's right now I'm not sure you've got other initiatives things like the open banking working group who are working as it happens on things like api's the the complexity industry is such that you know that what the open banking working group is doing it's got nothing to do with PST d2 because it was a work stream started before PST to hit the banks but then you can see how these things potentially could converge in the same way that the EDM council is doing work around ontology Zitz it's interesting as the Chinese as the Chinese would say we always got away with it potentially potentially it could change a lot and potentially will remain is an interesting discussion and keeps going on somebody somewhere along the line I mean it's either at the facto standard because the businesses adopted and become involved in it or it is mandated through a government entity as a standard to do something as a pure technology and technology company is pushing it or playing with it or proposing it isn't going to generate enough momentum for it to be disruptive to anything that its bank payments Swift or silos activities and so forth but it is a technology that's going to find its way in the next five to ten years into a lot of the things that transmit information about transactions information about exchange of physical goods some people are saying it will be the new barter economy and we don't need money anymore and I can't figure out how that happens but there are people who write about those things so interest level will be high but I don't think unless one of the two things happen either companies actually invest in it for their own reasons in a major way yeah well but whoever's looking at it typically whoever's looking it typically has the word innovation in their job title and that gives you clue the only people making money out of blockchain right now the conference organizers you know and potentially some of the big consultancies I mean and the real key thing is what are the use cases and there's lots of things you could potentially implement on blockchain right what are the use cases where blockchain will bring real return on investments and what real value add I know one a potential use case and one particular Bank is looking at it is trade finance now that is a very very complex process right now which needs digitizing and now could be one use case where blockchain would bring some real benefits to the banks does it bring benefit into the payment space I don't know I mean why you've got technology that does that today quite happily what benefits would launch it you could do it on blockchain potentially but what benefits would blockchain bring so these are the things that the banks the banks are looking at right now I mean we Intel has a sort of an interesting we've just released an open sore a lock chain implementation to the open source community and just it's not a product it's not selling we're not selling it we just gave it to the open source community the things that we look at are things around security and privacy scalability and performance and what we did a few things with a blockchain implementation that takes advantage of components down at the silicon level that help with things like security so a little bit of Intel stickiness there but it's partly torso you know so we can have a conversation banks and begin to understand but you know what the banks could potentially do but for a you know company like Intel this can be a very time-consuming conversation with the banks and lead nowhere so you know well the regulators are looking I mean I know so so we're talking about regulates and technologists of the the FCA over here has created an innovation sandbox and I know that there are some companies trying to get into that sandbox with blockchain technology so that the regulator's themselves are looking we know that Swift is looking at it because in a tribe has a complete you know stream are looking at it but it's it's mainly exploratory at the moment the I mean you mentioned trade finance for law enforcement could conceivably be interested for purposes of preventing mon
y laundering illicit gains in fact them and they may actually say any cross-border transaction will be executed through blockchain and we will have a virtual general ledger that Vesey regardless of who you are and that will be visible to us every step of the way and if you're not complying this automatically you'll be investigated because the assumption is your illicit transaction going on that's been in discussion between Justice Department in the Treasury Department to in the Drug Enforcement Agency in the u.s. is all of a sudden excited about this thing so all those things may actually make it more sticky and actually succeed but you need something like that for it to actually become real rather than on the investigation exploration to processing to automatically set the nice day we're back from that and say if I'm gonna trust machines to do that I want the most complete and expressive view of my data possible and so yeah it's something like marklogic where you can have a very expressive view relationships between is a good place to start to say I'm gonna send up a load somewhere and so machine's going to say yes or no to it I don't want to be constrained by it's a you know it's addition and subtraction when it comes to Bitcoin but when you talk about something like derivative it's a more complex it's actually calculus so if you want to represent it representing that is a bit more complex so so who knows maybe there will be representations of including the express inside of a payload that says yeah this particular instrument has about Providence that can tell the machine yeah this was derived from all of these underlying to this reference data somewhere over here my desire so if you want to command which happened sort of supernatural mistake the only way to do it is you out the economic transactions that mentions the previous one which is not very efficient so there are cases like regular three wise were not being sort of angles okay okay thing else he was asked to review a report of Deutsche Bank avoid ever better training for their blog chain across their investment bank and he actually calculated they would use half the world's electricity so many machines okay okay all right don't I know thanks thank you
