 Welcome everybody Today I'm going to talk about softcom, which is a collection of solutions and technologies That will drive the future of telecom in the next ten years So the title of the presentation is to reshape the future way of telco industry based on open stack I guess today every business is becoming a digital business Customer engagement models are changing all the time. We're now looking at more customer experience centric As opposed to KPI based in the past So where will things go operators are unsure of where things are going So they have to hedge their bets and basically adopting an anti-fragility model going moving away from stability so essentially In the past couple of years. We've seen saturated the market has been saturated and declining We're seeing declining our booze revenues We're seeing erosion of revenues from OTT players Voice being replaced by Skype and SMS by what's up and the like various OTT applications We're also seeing saturation of markets, particularly in developed countries in Western Europe So penetration levels are gone well over a hundred percent So there's not much more growth except when we talk about M2M And in the face of this traffic is increasing Which are which are leading to a lot more costs and capex to deal with this traffic going forward so this model is unsustainable and We now move into a future model So what role can the telco play in the future model in all of these various networks? We have here It can offer virtual networks. It can offer platforms for M2M markets. It can also partner with OTT players It's also offering services today and competing with the likes of Amazon and Google offering Various things on as a service platform as a service infrastructure as a service So Then we the bottom left hand corner here. We're looking at the own the cost structure the total TCO or total cost Total cost of operation which basically is the capex and the OPEX So for an operator that has built out his network Typically most of the cost is coming from OPEX and this is in has been increasing as we can see here It's not very clear but in 2002 the OPEX constituted 77 percent of The overall TCO and that's now moved up to 82 percent So this is increasing all the times and is unsustainable So needless to stay operators are now looking at ways to reduce the costs and they're looking at technologies like NFV etc Then we move over to the the future cost structure, which has to be more optimized more flexible And basically less cost We need to have much more efficient network management And we need to be offering more customer services more personalized services And this will drive innovation So we're looking to borrow technologies and adopt technologies from the cloud IT We're also adopting new methodologies of trying to reduce the time to market for new services and doing development and operations in parallel because DevOps and This is a model that's being used by OTT's So in the past we may have take is taken like a large tier one operator Maybe up to one to two years to develop and roll out a new service now. We're looking at a matter of months weeks and even days We even talk about the capability to offer say a more a mobile virtual network operator in the matter of five seconds five minutes Sorry, so that those capabilities are now there So here we look at where we have been in the past ten years and how we've moved on So in the past we've had all these legacy networks had siloed They were not layered with a dedicated network for a dedicated purpose and over the past ten years TDM ATM SDH networks have now been all replaced by IP Everything is becoming all IP or is all in most cases most networks are all IP today And we also see Single strategy so instead of having multiple layers and multiple boxes We have for example a single ran one one blade server can handle all technologies 2g 3g and LTE with a single core network single EPC of all packet core and Everything has been simplified What moving forward We're now adopting the technologies into the telco space from IT and cloud So we're looking at software defined virtualization automation distribution intelligence and so on and This is what will reach shape Next test ten years. So collectively all these technologies. We're calling softcom which is stands for software defined telecom So here again, we mentioned the challenges We know what the operator response needs to be he needs to simplify its network going forward It needs to become more operationally efficient and it needs to monetize new services and create new revenue in face of declining revenues on The right-hand side here. We see there are four main pillars within softcom in the last presentation I think we heard about SDN and NFE SDN software defined networks is basically Decoupling the control plane which controls the traffic from the user plane It's an abstraction of the transport layer and Basically allows for more programmability of the network and a more optimized Configuration in real-time to be able to deal with massive loads of traffic and you know, there are some challenges with this technology One of the main challenge I guess the moment is that it's still under standardization Not all the standards are there for it yet. Well, there are some initiatives like open daylight So this needs to move forward. There's a lack of expertise in this area There's also it's not completely Clear how Operators will evolve and migrate these technologies which which areas they should focus on should be service chaining first should it be the IP optical layer So obviously there's a cost-benefit trade-off which needs to be proven then we have Moving on then but of course the main driver here is to be able to have a more programmable network easier to provision and also be able to manage traffic flows better and Also to be able to monetize new services. This also gives us that capability The second technology here we're looking at we've mentioned in the last presentation was a network function Virtualization and here we're talking about decoupling hardware and software so, you know in traditional networks today whether it's the you know, the EPC or the various components of the network you have All of the functions are made by one vendor all the services of it is integrated It's a one-box solution Based on at car hardware now. We're looking at this obviously a way to reduce costs is to be able to use standard volume-based industry commodity hardware and servers storage and networking equipment So that's this basic the main advantage of course with NFV and this is getting a lot of momentum now Probably more so than SDN and there's a lot of trials and commercialization starting in this area is Is the is the cost savings and the cost savings mainly would be in The the operational side of things so you can have a more scalable more easy pool of resources And you can buy them and mix and match them the whatever way you want You don't have to buy everything from Hawaii. You can buy part of it or you can buy all from one vendor whatever way you like And we're looking at virtualizing mainly I guess Probably the core network. We're also looking at virtualizing the IT environment that an operator would have as well so that moves us on to the Top right-hand box of the right-hand corner the IT cloud So here we're looking at the third pillar of softcom which is basically Taking all the IT infrastructure business support systems the billing systems ERP You name it just maybe 10 15 front-end and back-end systems and basically Putting into a virtual environment virtual data center Again to get the economies of say scale the capacity and the scalability and be able to scale You will save a lot of money. This is a very obvious area for because it's already kind of like a data center And then the fourth pillar is what we look at is Internet like operations so and here we're looking at bringing more personalization more services where the user can create its own Services from a web-based platform It's becoming more like a Google or is Amazon. It's also being able to offer new types of services For example traditionally today, we have infrastructure service platform as a service software as a service now in the future We can offer a lot more types of Variants Here we have a network as a service. We can offer network as a service could be an MVNO. It could be VPNs Then we have data center as a service as well on the next slide. So these are the four pillars of softcom This is just some of the some of the solutions and basically the hardware of the data center portfolio from Hawaii and Basically, this is existing to Hawaii is an IT enterprise supplier today provides all these different solutions for network server storage and data center But in the telco environment, we'll be using for example, we can offer the fusion engine, which is a carrier grade The cloud engine and we also then have the the e9000 blade server, which is Also used a fusion cubes also used for for the softcom solutions Yeah, this is just more details in the hardware What what is required for the telco environment is you know, we need things enhancements over and above what we get from the IT world We need like for example more resilience more availability reliability 800 millimeters depth nebs compliancy something that's very important also then the ability to support both DC and AC Because if you're swapping out some many exchanges only support DC so you need to be able to support that as well as AC which is trick traditionally the IT type of power supply and different sizes of equipment as well So Engagements, this is for NFV. We have we've started engagements in 2012 and it's ongoing today So we're basically these are some some examples here where we're doing virtualize Virtualization of different functions. IMS sits in the service cloud IP multimedia system And this is one of the most popular things at the moment primarily for delivering the voice over LTE services We're actually running into commercial trials on that at the moment So cloud edge this is something specific to the carrier world Typically the The data center you're talking about consolidation and centralization of everything However, for the carrier world, we need to move into a more distributed environment In some cases we need to be closer to subscribers closer to the base stations to deliver the traffic So the cloud edge is is designed for deployment Distributed type of deployment As high resilience and high availability So we get carry great performance without sacrificing performance Right now in this particular environment, we have various functions that we've Virtualizing which is the valve packet core which is the core network for LTE Multi-service engine we have the mano which is the Management and network orchestration for the virtual machines. It's part of the Etsy Definition and here we can see fusion sphere, which is the cloud OS that for Hawaii on top of that sits the infrastructure so this high-level just overview of let's say How this will look like we have three let's say layers of clouds we have You know, there's the service cloud which is where all the applications will be hosted We have the operation cloud just taking care of OSS and mano and we have the control cloud which is networking The data center and data center interconnection And then as I mentioned before we had in the middle We have the cloud edge and on the far left-hand side We have the cloud baseband the cloud baseband unit So different levels of cloud here So this differs from the it cloud because the cloud edge for example is is not the endpoint traffic will will not end there So on the right-hand side here we have in the service cloud we have for example the video content delivery We have the IMS and the operation cloud we have the NFV Management and network orchestration and the cloud service broker and as I mentioned before Control cloud would be the data center SDN controller and the day center interconnection SDN controller in the middle then we have the What's called a software-defined edge? And this is where you have let's say the metro or the aggregation in the traditional network And we have different components here for example We would have what's called a single radio Network controller It's funny because in in in LT we got rid of the the radio network control controller the RNC Which was part of 3g but now we're bringing it back here Because we're trying to manage traffic flows through three different types of of of networks 2g 3g LT maybe LT advanced But also Different types of cells. We have small cells macro cells Wi-Fi Headnet so-called and to manage all that we need we need to have the single radio controller Another important area here, which we're doing quite a few trials with is is service chaining the color here the GI LAN chaining which is It's behind the packet gateway And here we're looking at You Optimizing and using resources only when they are needed so we don't need to you know have all resources for every single session for every single call whatever to give an example here we we If you have let's say a customer who has used up his his allowance for traffic for the month and you have to throttle them back Then there's no point in putting in video optimization video optimization equipment into that Service flow so you take that out so basically it simplifies things and it also It's quicker to develop to launch new services as well. It's actually becoming a very important feature And here also we have The content delivery network can be virtualized This can also actually Be also in the service cloud it can vary in time of day We can move this content delivery both backwards and forwards Or if there's faults in the network and the session border controller as well another another thing that's not mentioned here is Multiservice edge and this is basically Like making pushing the traffic further down towards a subscriber making the base stations more intelligent and being able to offer more personalized services for example, you do more Active device tracking location You could do augmented reality Content delivery You can do video analytics All sorts of new services and this is a new initiative that started Which Huawei is backing as well So then we have the software defined baseband, which is we're now in the access part of the network Here we have a combined 2g 3g LTE all in one we centralized control and we have a clustered network So typically before you'd have a base station would sit on One single site and now we split up the the RAU which is a radio resource unit This now sits on the antennas or the mast and then Baseband, you know would have been in a shelter for example But now this is all moving back to say to a transmission have been aggregated and the only problem with this is You need really need to have fiber out to your base stations Because of all the signaling and stuff Finally, we have the software to find radio frequency Here we're talking about multi-band different many many different bands that need to be supported Multi-mode multi-generation all in one blade and this exists today In terms of Migrating your networks typically what's happening today is usually we're starting on the right-hand side so most of the the interests and most of the Say let's say the virtualizations that are going on are happening in the In the service cloud on the right-hand side, but we're also having initiatives then in the in the In the cloud edge So this is the mobile broadband network and then it's quite similar in the in the fixed broadband network Again in the service cloud we have The IP policy stn controller. We have the DC interconnect stn controller In the cloud edge. We can virtualize The set top boxes the CPEs and the access routers access routers would be for For the enterprise and we've actually we're actually working with one major tier one operator in Europe at the moment and doing a Trial on this and also it's looking very positive from cost savings the ability to be able to instead of before you had to Manually update all the firmware or it now you can do it all centrally from the software to find edge You've also got the VDI, which is the virtual desktop infrastructure you can offer Again, you have the content network and the and the border network gateway Optical and IPS dn controller This this is a Basically something new where basically the there that the centralized controller will have a view of the network He will know which OTN box The routers are connected to will have a view of the traffic And he'd be able to create traffic engineering tunnels to route traffic at a particular time when there's a problem in the network or there's a Big demand he can create this to one particular destination But by controlling at the same time the MPLS routers the PPP and PE routers But also the the OTN WDM network, so it's basically a high-level view That you could not do today with today's network And then finally you have software to find access Which is basically the operation and maintenance automation of the optical line terminals In the case of fiber to the home or fiber and then MXU, which is also can be virtualized for DSL and Before we saw some NFV engagements now. I'm just looking at the soft-com engagements. So we have about 50 plus Joint innovation projects with 20 leading operators around the world So basically we're involved with every major operator at the moment and seven of those projects are already in commercial use as Most of these here would be Either NFV or SDN. I think you're looking at SDN. You're looking at The transport SDN TSDN the STIP ran the IP core traffic optimization and Yeah, that's a service chaining the rest of NFV But we're also doing trials on the other pillars of soft-com as well so bringing it back then to How it looks in terms of the various stacks and how it relates to open stack We have this is a cloud DC infrastructure So fusion sphere is the name of our cloud OS, which is based on open stack and It provides Extra in addition to open stack. We've got these APIs North northbound and we're also we have the standard ones then southbound so The different the add-on here again, we're talking about there's a carry enhancement where we We're looking at much more distributed environment than the IT world. We're looking at higher availability lower latency and migration in Terms of the overall portfolio within Hawaii. I think we've got about 120 deployments of this 500,000 virtual machines So we have support for SDN NFE initiatives about public and private cloud services and data center consolidation so what is How are we in a v proposition essentially Hawaii today can do a complete end-to-end solution As we saw before we have our own hardware We have our own cloud platform which is developed and based on open stack We have our own applications for the mobile for the fixed We have our own with the virtual infrastructure the network services and Also now we're positioning ourselves to move in as a system integrator We've actually got a project ongoing now at the moment for delivering a voice over LT where we're playing the role of a system integrator So essentially everything we can provide if required Again is another look at how it relates to open stack I'm only showing here three of the main Modules Nova for compute neutron for networking sender for storage There's obviously about five more different modules that we're using it at the moment So what you see here in in the orange color is what's being developed by By Hawaii, so we do have our own hypervisor, but we can also inter work with the standard hypervisor like KVM We're developing our own API extensions as well And then we have a thing called fusion manager, which is managing the life cycle and user management user authentication So some of the some of the additional things we're talking about with API's carry grade hypervisor We have an SCN overlay network and we've an ultra large storage pool for IO acceleration These are some of the initiatives we're working with in some of the standardization groups Obviously, we've been working with the open networking foundation for a while We are a chair of the migration work group and we're also engaging in harmonizing efforts with the IETF and the broadband forum In Etsy we act we have two major roles We have co-chair of the working group for architecture for virtualization infrastructure, and we're also acting as program manager Then of course on the right-hand side, we were a silver member of open daylight Which is involved with promoting and developing the STN controllers and last but not least We are a gold member of open stack. If not, I think we're one of the Top ten in the top ten contributors to open stack This is the last slide in my presentation. It's just basically again another initiative which was mentioned in the last Presentation as well where we're talking about the open platform for NFV, OP-NFV and Again, we're a platinum member in that There are 39 members in this and the object of this initiative is to accelerate NFV And I guess a part of the problem is that this is a very complicated Complicated architecture. There are lots of interfaces and lots of interfaces to be defined And what is the best way to manage and control that so the initial focus of open NFV? You can see there would be in the red box Basically dealing with the the virtual computing storage network the hardware and also the infrastructure management So that's it. It's quite short Thank you, and if you have any questions, I'd be happy to answer them. No questions