Engineer UC voip

Job Specification VoIP Carrier Pre-Sales Engineer

VoIP Carrier pre-sales engineer

Netaxis are a rapidly growing Telecoms Systems Integrator with their own in house software development team. Partner solutions from Oracle, Genband, Sonus, Broadsoft etc are combined with our own suite of products that include Network Monitoring, Call simulation, anti-fraud, Session Routing Engines, API Orchestration, Video and WebRTC solutions.

In their Benelux home territories Netaxis sell to most Tier 1 and 2 telcos and the company is now successfully expanding in other European markets including France, Italy and the UK and also around the globe with customers in Asia and Africa.

The company is now looking for a UK based Senior pre-sales engineer with knowledge of the UK market to assist with the business development process. Whilst this is a technical pre-sales role the successful candidate could also be required, at least in the short term, to assist with Systems Integration activities. The job will also involve overseas travel, mostly to Europe but there is the potential for the occasional trip further afield.

Skills required include knowledge of the design and deployment of Session Border Controller and Softswitch platforms together with the accompanying suite of products and services. The candidate must be a native English speaker and be a good communicator. Key relationships will be with vendor partners, customers and your own colleagues across Europe.

Drop me a line if you are interested and think you fit the bill – tref at netaxis dot solutions. The salary and package will be whatever is right for the individual concerned.

I am not currently seeking referrals from Recruitment Companies.

Engineer fun stuff peering

UK & Ireland Peering Forum Social

Informal social event for Network Engineers and Peering coordinators during London Technology Week – UK & I PF Social

Once again LONAP is joining forces with the other mutual IXPs in the UK and Ireland to promote Peering and facilitate the exchange of ideas in an informal atmosphere. Attending this event will enable you to network and engage in technical discussions with participants of all the major UK and Ireland Internet exchange points.

This event is specifically for Network Engineers, Peering coordinators and anyone who has organisational responsibility for arranging settlement free interconnection across one of the organising IXs.

This informal Social event is being held in London during The London Technology Week on the evening of the 12th June, address and registration link below.

Please do register early as places are limited.

I look forward to seeing you on the 12th June.

PS I’ve had lunch a the the Fellow, the venue for this UK & Ireland Peering Forum Social. It is very convivial and we are guaranteed to have a pleasant and useful networking evening. The pub is on York Road down the side of Kings Cross Station.

Loads of other peering posts on this blog. LONAP btw is  a ‘not for profit’ Layer 2 Internet Exchange Point (IXP) based in London. Our data-centres host a network of interconnected switches providing free-flowing peering to help minimise interconnection costs. We provide exclusive connectivity between members, who are effectively LONAP stakeholders. This ensures that LONAP members enjoy excellent value and maximum benefits.

Traffic exchanged between LONAP members, reduces volumes sent through upstream providers, reduces IP transit cost and bandwidth usage. Our membership includes ISPs, network operators and content providers with their own data networks. We provide regular opportunities for members to network and meet new suppliers, and support operators in growing their portfolio and reselling LONAP connectivity to networks outside of London.

Engineer voip

Looking for a Session Routing Engine? It needs to be carrier grade

Carrier grade architecture for Session Routing Engines

It’s a fact that many ITSPs, certainly in the UK where there is a proliferation of them, do not need a call routing engine. Their (mostly) least cost routing needs are covered by their existing softswitch or by the Session Border Controllers they stick in front if it. As you get bigger however you do get to want more sophistication in a Session Routing Engine.

For example if you want to add more logic to the routing decision. I’ve already written about routing based on MOS scores but there are a whole load of other scenarios where routing decisions may be influenced by external factors. For example in scenarios where users are switching between persona and want different voicemail policies to be applied between personal and business calls.

The Netaxis environment supports this type of scenario and our Session Routing Engine plays an important part in it.

SRE was built for Tier 1 and 2 operators who need sophisticated routing capabilities. As a relatively new product we have also been able to play in features that make a carrier’s life a lot easier.

SRE Routing Framework

The SRE routing framework guiding principles are flexibility and openness. The data model and routing logic let you have unlimited routing policies.  Call handling is defined as a set of sequential actions – the SRE comes with a rich set of predefined actions that drive the  creation of the routing logic.  New actions can easily be added to build up fully customized routing logic.

In addition, SRE is natively designed to query third party elements. This ranges from simple queries such as external database interrogation to much more complex scenarios like PBX interaction.

Service Logic Editor

The Service Logic Editor has extremely powerful and versatile functionality and is designed to let you create service logic in a few clicks. It allows the creation of any kind of route including those that are perhaps very specific or temporary/transient. Routes can be run in simulation/validation mode before being deployed to the operational network.

Carrier Grade & Scalable

In considering the architecture it has also been important to recognise that the end product has to be carrier grade. Resilience and capacity scalability is key.

An SRE based solution can be made of several Call Processing instances managed by a centralized Element Manager. The multi-instance approach offers wide deployment possibilities and huge scalability. Load balancers can be added to complete the picture. Call Processing is fully redundant working in High Availability mode.   

Long experience has also told us that things can go wrong when rolling out network upgrades, even if you have tested it to the umpteenth degree. Bearing this in mind the SRE has a really cute data versioning feature. The SRE always keeps 2 set of configuration settings that can be toggled with a clock of the metaphorical button. Reversion to a stable release in the event of a problem during an upgrade is easy.

Who could ask for more? 🙂

If you do have any questions or would be interested in finding out more about our Session Routing Engine feel free to let me have your name and work email  in the form below and I’d drop you some info. I won’t spam you but I need to know I’m dealing with serious people so no generic gmail type addresses please.

    Netaxis is a competence centre for telecoms and now operates in the UK 🙂

    Engineer internet ipv6

    Transfers of IPv4 Addresses from /8s Held by One Organization

    Guest post on IPv4 Address market by Sandra Brown

    sandra brownSome interesting insights into the IPv4 address market by Sandra Brown who has been working this space for a number of years now. Read on.

    In 2011, as Nortel’s Director of Engineering I started the orchestration of the sale of Nortel’s  Over time this undertaking has resulted in the sale of approximately 85% of the /8.  Something that has been discussed in the ensuing six years, is how many other “organization” or “company” owned /8s might ultimately be sold.  IPv4 Market Group has examined the before and after records in the Regional Internet Registry (RIR) databases, in addition to the RIR transfer logs, to see how many of these /8s have moved.  All data being analyzed is publicly available.

    As shown in Figure 1 below, there are 26 /8s that fall into the category of “organization” held.  In this inventory, we ignore /8 blocks held by the US Defense Organizations, because there is no indication that the US Government has any intention of moving them in the future.  We do include the UK Government’s two /8s, as we have seen past movement by the Department of Work and Pensions.   We also include the US Post Office, as it seems likely this organization might desire a few million extra dollars at some point.

    ipv4 address market

    Figure 1: List of 26 /8s – organization or company owned

    We are seeing that 10 out of 26 entities have “sold” a portion of their company held /8.  These entities are shown in Figure 2 below.  As of May 9 2017, 1076 /16s have been sold from the aforementioned /8s.  This implies that approximately 16.2% of the 26 entity held /8s have been transferred.

    ipv4 address market

    Figure 2

    Some of the transfers from these ranges, including other large transfers not from a /8, are shown in Figure 3 below.

    ipv4 address market

    Figure 3

    There are notable points to be made:

    1. We know that approximately 16.2% of the 26 /8s have been transferred. This means that organizations are interested in monetization and that the trend should continue as prices rise. 
    2. Most of the transfers from the /8s have been large transfers of /13 to /10. Big block holders are not selling their /8s one /16 at a time! (this is leading to a shortage in the /16 market) 
    3. Can we predict how many of these /8s might still transfer? Those entities that have already sold some of their /8 range will probably continue to sell more over time.  While 42% of the 10 /8s have already been sold, perhaps another 40% will move in the next five years, totaling 1024 /16s moved from these entities. 
    4. What about the 16 /8 blocks from entities that have not yet sold a /16? We can speculate that some companies will never sell because they don’t need the money, and having the IPs is more of a strategic advantage than selling them.  Or, they might not sell because even though the IPs may be sparsely used, it would take more work than they are prepared to undertake to free the blocks for sale. These could be the reasons why Apple, HP, and IBM never sell. However, other /8s may be available.  We can scan the list and predict that at least six of these /8s will be partially sold in the near future.  If they sell half of their holdings, it would total another 768 /16s in the IPv4 transfer market. 
    5. Who is buying the large blocks? There are 14 different buyers shown in Figure 3  Two entities in particular, Microsoft and Amazon, have received large portions of the transferred IPs.  Microsoft has bought from 4 of the 10 sellers (CSC, DuPont, Xerox, DWP) and Amazon from 5 of the 10 (Merit, MIT, Xerox, DuPont, Merck). 
    6. Is there a market for the large blocks of the future? IPv6 is still years away from a critical mass and we continue to hear about the costs and headaches of its implementation.  As long as IPv4 is the “best” and cheapest solution, we will continue to see a market.  The question then becomes, “what is the expected price for these large blocks?”  We are already seeing $13 to $15 per IP for large holdings, most likely because they are so rare.  Is there a price point where buyers walk away?  Yes, of course, as we are seeing some companies refuse to pay the market price already.  Microsoft and Amazon have deep pockets, so it would be speculation to put a number on their maximum.

    This analysis of the 26 single-owner /8s has shown that these large blocks are being transferred in large sub-allocations, and the /8s are not filling the /16 market.  Ten of the 26 have already done some selling, and we expect this trend to continue, to more than double the large block sales made so far.  The IPv4 Market has some legs still!

    Sandra Brown, IPv4 Market Group

    Engineer voip

    MOS based route selection using Netaxis Session Routing Engine

    Session Routing Engine is us

    Had a really good working lunch last week with special guest Laurent Debacker. Laurent runs fixed and mobile network operations for Belgian telco Voo and provided us with a very interesting talk about their implementation of the Netaxis Session Routing Engine.

    Some of you might think there is nothing much to say about routing engines. After all, you just set one up, point traffic in the desired direction and your Uncle’s name is Bob. Maybe you employ least cost routing and periodically upload new rate cards. That’s about it when it comes to routing engines. Isn’t it?

    Well maybe it is and maybe it isn’t. Classically routing engines are unwieldy beasts and offer little incentive for people to do anything imaginative with them. However the telco world is rapidly evolving and your routing implementation needs to evolve with it.

    Here I offer you one scenario:

    Least cost routing is important because if you don’t watch your supplier rate cards you can quickly rack up losses on any given routes. However if you don’t watch the quality of your chosen routes you can quickly lose customers, or at least build up sufficient levels of disatisfaction that they start to bombard your helpdesk.

    If you had a means of actively monitoring the quality of lines to specific destinations you could avoid this. Happy customers mean higher spending customers.

    Netaxis have a suit of products that are built around our Session Routing Engine. The combined system includes the Nemo Network Monitoring tool which provides MOS scores for specific calls and the Dory call simulator. When combined with the SRE you can produce data on the quality of specific connectivity partners and make routing changes in real time based on this quality. It is also possible to get network alerts telling you that specific destinations have had changes to the carrier. The new carrier may not be the lowest cost but could serve as an interim solution whilst you address the quality issue with the original.  

    If you want to know more about how we can do this just get in touch.

    Engineer peering

    RIPE74 is in Budapest

    RIPE74 report from Budapest

    Sat in the RIPE74 meeting having arrived at tea time last night. Every time I come to these meetings I am reminded at how good these events are. It’s not just the content. It’s the opportunity to meet folk relevant to your business that you wouldn’t normally see on a day to day basis. Here they are all in one spot and you spend your time catching up with them and their business.

    These meetings – RIPE and their ilk – are how LONAP do our marketing. We are a bunch of engineers rather than sales types. We chat with engineers about engineering at engineering events. LONAP is an Internet Exchange Point for engineers.

    RIPE74 happens to be in Budapest. I’m a Budapest virgin. It looks an impressive place.  At first glance. The featured image is the view from my room. I’ve not really had a chance to get out of the hotel yet although Richard Irving and I did get out for a quick bite around the corner last night. The opportunity will come. RIPE74 finishes Friday and I don’t fly home until Saturday lunchtime.

    I’ll be sticking more pics from Budapest up on Facebook as and when I get the chance to take some. Budapest, I found out from the taxi driver en route from the airport is divided into two halves separated by the river Danube, Buda is the historic area seen in the pic and pest is the newer less interesting bit. Apparently. Famous for it’s outdoor spas. Will have to see if we get a chance to go to one. The days are pretty busy.

    If you want to chat about peering at LONAP just tap one of us on the shoulder. We are proudly wearing our black LONAP polo shirts. Loads more peering posts on this blog by the way. Check ’em out.

    Engineer engineering voip

    Self Learning Control Layer in VoIP Networks

    Taking Control of Your Network – The Self Learning Control Layer

    The development and implementation of telecommunications networks and services has traditionally been a long drawn out affair. Even the smallest project might have an 18 month gestation period. Major network roll outs much longer. The advent of API driven environments provides us with the opportunity to change all this and here I discuss the concept of the Self Learning Control Layer (SLCL). The SLCL is an intelligent overlays  to VoIP networks driving efficiency, reducing costs and making it easier to add new products and services.

    Self Learning Control Layer Building Blocks

    An SLCL has three building blocks:

    • Data Mining,
    • Call Control &
    • Orchestration.

    Data mining is used to monitor activity and health in a network. Call Control is used to effect change to traffic routing based on information provided by the Data Mining and Orchestration. Orchestration makes provisioning decisions based on predefined policy, inputs from both the network and from end users.

    The types of learning the network might make include:

    • addition of capacity to or traffic rerouting away from congested areas – eg based on MOS scores
    • automatic shut off of calls considered to be fraudulent traffic
    • Routing policies altered by a change in an user’s status (eg from “business” mode to “personal”) or the provisioning of additional capacity based on the addition of new users to a network.

    The SLCL is designed to make your network easier and cheaper to operate to make your customers happier. Being API based it also makes it easy to integrate with your existing network infrastructure. The SLCL is a concept developed by Netaxis who provide all the building blocks you need to implement it.

    I’ve already discussed elements of the SLCL in my posts and workshops on VoIP Network Monitoring, fraud detection and call simulation.

    Session Routing Engines

    My next focus is going  to be on Session Routing Engines. The SRE is a key component of the SLCL and is the heart of the Call Control function.

    Essential features include:

    • APIs to standard network elements such as softswitches, Session Border Controllers, VoLTE and IMS and indeed other existing Core Routing Engines
    • SIP & ENUM
    • Be flexible and easy to implement

    The SRE is the subject matter of my next working lunch which is in London on Wednesday 3rd May. At this event we have a special guest, Laurent Debacker of Belgian telco Voo who is going to talk us through their own implementation of the SRE and act as a catalyst for a general discussion around this subject.

    You can check out details of the event here. It’s free to come to but you have to be a VOIP network geek for it to be worthwhile for all attendees – we want an active discussion on this subject and places are limited.

    Engineer voip

    VoIP Network Monitoring Workshop Wrapup

    VoIP Network Monitoring workshop learnings

    Terrific VoIP Network Monitoring workshop on Tuesday. Ended up with around 25 engineers in the room. Who would have thought that 3 hours would whizz by when talking about this subject? Thanks to all who came especially Raph from Netaxis, Ade and Ben from Zen Internet and Florian from Oracle. They all did a sterling performance with their respective demos.

    These small workshops are great. You always learn something new and there are a lot of nuances to this subject. I found a particularly useful feature on the Netaxis Nemo tool (OK I am with Netaxis but it is an useful feature:) ) Nemo can correlate CDR data with the traces and logs of the call being made. It is therefore very easy to track down a recording that may have been made using a network probe based on the number of the calling party. Handy in situations such as call centres where someone may ring back and complain that what they got was not what they asked for previously.

    I think the biggest issue this industry is going to have with VoIP Network Monitoring is in the move to the cloud. Currently for anything more than a low number of sessions it isn’t particularly practical to monitor RTP traffic in the cloud. Data throughputs mean you need physical network TAPs to mirror the traffic and the cloud ain’t physical is it. Picture the scene. A massive AWS datacentre with rows and rows of identical cabinets but with one TAP sticking out like a sore thumb.

    If all you want is signalling data you will probably be ok but most network managers want to be able to listen to real media when troubleshooting problems.

    Afterwards we all trooped to Craft Beer on Mitre Street and then on to The Lahore Kebab House on Commercial Road. This curry house is a real find. Great food. No booze but you can bring your own from the off license next door. I especially recommend the chicken tikka and lamb chops. If you’re ordering onion bhaji that’s a meal in itself.

    I digress. Get in touch if you want to know more about what we can offer in the way of VoIP Network Monitoring. Or anything else in the VoIP network game really 🙂

    Next workshop is about Session Routing Engines on Wednesday 26th April. Exciting eh? Save the date. More deets to follow.

    Thanks to Oracle and Netaxis for the food and drink afterwards 🙂

    Engineer voip

    More VoIP Network Monitoring Stuff

    All about VoIP Network Monitoring

    The UK VoIP industry is maturing. Networks are growing, subscriber numbers are increasing and the tools people used when they first started may no longer do the job.

    Once upon a time it was just fine to use freeware. After all there is plenty of support out there. Usergroups willing and eager to help. On a best efforts basis obvs, or paid for even. Now you haven’t got the time to do all that stuff yourself nor all the learning that has to accompany it. There comes a time where the growing ITSP that once managed on a shoestring has to face reality and begin to look at turnkey vendor solutions that come with their own support and carrier strength operations.

    Network monitoring is one such area. I’m finding a real interest in this subject in the UK. My VoIP Network Monitoring Workshop has been very well subscribed. Of course it helps that there is free beer and curry afterwards but to be honest that is by the by. The UK tech community is interested.

    I wear a number of hats where VoIP Network Monitoring is concerned. Netaxis have our own product, NEMO, which has just undergone its third major release. Netaxis also act as a Systems Integrator  and support a wide range of 3rd Party vendor tools. Included in this is Oracle’s OCOM integrates well with Oracle’s own Session Border Controller line as well as via external probes.

    It’s only recently I’ve really dug into this whole subject. Obviously things do go wrong in any network and especially in VoIP network where there can be so many topologies involved. Also as your number of subscribers increases the more likely it is that you will encounter problems. Because voice is mission critical to a business these problems are going to make for uncomfortable times unless you have the right troubleshooting tools.

    This means being able to drill down into into individual voice calls to see what is going on. What’s the voice quality (MOS Score) looking like? Is the problem only happening at particular times of the day? Why is that phone constantly re-registering? Is it a site wide problem? Are calls timing out after the same amount of time? Is it handset specific? Codec specific? Lots of variables to contend with.

    When your network grows you also need to start thinking about how and where you deploy your VoIP Network monitoring tools. Network data needs collecting, storing and analyzing. How do you work out how much storage you need? What’s the profile of your customer base? How many simultaneous calls? What’s the average duration? Which calls do you want to record? All of them or just ones that you are trying to troubleshoot? How long do you want to store the RTAP data for.

    All of a sudden you are into an engineering effort just to figure out what the specs of your probes are going to look like. Will you need much bandwidth between the probe and your network monitoring server. What TAP spec do you need? Traffic has to be mirrored somehow.

    There’s a lot more to VoIP Network Monitoring that people sometimes realise. What’s more the scope of what is wanted out of a VoIP Network Monitoring tool has grown. Firstly it’s a natural source of data for call recording purposes other than mere troubleshooting. Regulatory requirements perhaps?

    Then there’s fraud. If you are monitoring all your network traffic it seems to make sense to be able to analyze this traffic for patterns that suggest fraudulent activity. What steps do you take for fraud mitigation? Does your VoIP Network Monitoring tool also help you detect fraud?

    I think I’ve gone on long enough. If you want to join the discussion on this subject just sign up for my VoIP Network Monitoring Workshop on 21st February. Don’t come along just for the beer and curry though. I’m after participation. Involvement. 🙂

    Oh and if you want to talk to me directly about what we can do for you in respect of VoIP Network Monitoring just get in touch. Y’awl.

    Engineer voip

    VoIP Network Monitoring Workshop Update

    VoIP Network Monitoring Workshop update stuff

    Quick VoIP Network Monitoring Workshop update I have planned in town on Tuesday 21st February. The workshop is being held at Gamma’s offices near the Gherkin (St Mary Axe) with free beer and curry generously provided by Oracle and me, with my Netaxis Solutions UK hat on (blows on fingernails and polishes).

    I’ve booked the pub and the curry house. Pub is The Craft Beer Co at St Mary Axe where we have the downstairs room for a couple of hours after the workshop. 5 – 7pm say. Check out their beer . Some seriously good stuff there.

    Secondly I’ve booked Lahore Kebab House on Commercial Road. This is not only supposed to be one of London’s top Pakistani gaffs and can handle large groups but comes recommended by people I trust about this sort of thing.

    If you are in the industry and can’t make the workshop but fancy coming for a beer then you are welcome to do so. Just let me know. only beer though. Curry is reserved for the hard core who sign up for the VoIP Network Monitoring Workshop.

    Currently we have 24 people coming to the Workshop and have room for 30. More if you are late to the party but run a VoIP network and are really interested. Again let me know. At the moment I have 6 spaces left with  3 people who i know are coming but have yet to sign their names in the box.

    I have now run a number of workshops at the Gamma offices in the City. They have great meeting facilities and are very happy to loan me the space. However I have had bounceback from one or two network operators who are Gamma competitors. I am happy to move the show around if you want to get in touch about using your space. Needs to be able to cope with at least 30 and up to 50 people with some coffee facilities (beer is also acceptable). In return you get  some great exposure in the ITSP community. Let me know.

    That’s my VoIP Network Monitoring Workshop update. If you want to come sign here.

    Engineer peering

    LONAP traffic continues to grow with new peak

    LONAP traffic up and to the right

    LONAP traffic continues to grow with a new peak this month of over 240Gbps. I’d just been thinking of rattling off a quick post telling you we have broken through the 200Gigs mark at 206Gbps when lo and  behold that post became instantly obsolete.

    There have been 3 or 4 record levels of traffic in the last month and it can only go in one direction.

    Traffic growth is being helped by some of our large eyeball and content provider members who have been provisioning 100Gbps ports this year on our new Arista based core network.

    It’s such a pleasure to be involved with LONAP. A pleasure and a privilege. The team is great and the members are all great. Andy Will and I are going to be at RIPE73 in Madrid next week catching up with existing members and prospects. Look out for a little competition we are going to be running on Facebook.

    Don’t forget to sign up for our  Workshop and Christmas Party on 8th December. Password is lonap.

    For those unfamiliar with the Internet Exchange Point world The London Access Point [LONAP] was first established in 1997 as a ‘not for profit’ Internet Exchange Point for London. Today, our list of active members includes global brands, London businesses and FTSE100 companies, all of which are joint stakeholders in the organisation. As members, they all have exclusive interconnectivity and direct input into the configuration of the network and enjoy the commercial, operational and social benefits of working with a vibrant and dynamic exchange.

    Our growing membership includes ISPs, network operators and content providers with their own data networks. LONAP members exchange traffic using a network of interconnected switches hosted in our data centres across the City of London and Docklands.

    Ciao amigos. See you in Madrid?

    other peering posts on

    Engineer voip

    how do you perform interop testing after firmware upgrade?

    voip interop testing made easy?

    I’d be interested in chatting to anyone out there who has to periodically does voip interop testing to certify equipment after a firmware upgrade. You may be an IP telephony handset manufacturer or a vendor of other equipment that needs to be certified against an operator’s platform or network.

    This could be the subject of a future workshop. The whole process, which has traditionally been quite expensive because of the time it took to perform manual testing of sometimes hundreds of features is ripe for automation.

    If you use a voip call simulator such as Dory this is fairly straightforward but I wouldn’t mind hearing how others do it.

    Engineer voip

    What’s the Dory Story?

    Call simulator finds problems before your customer

    Dory call simulator can be used to quickly certify new firmware releases on CPE such as PBXs or to stress test a network after upgrades to carrier platforms.

    Some of you will know I’ve been doing some consultancy work for a Belgian outfit called Netaxis Solutions. Netaxis develop software that is used by telcos around the world to help run their networks.

    Recently I’ve been looking at their Dory call simulator. I hadn’t quite sussed how useful these thing are. If you are performing upgrades to your hosted platform then the ability to run stress tests overnight before your customers come in in the morning could save you a lot of angst. An upgrade that at first appearance seems to have been successful isn’t really tested until the network is carrying lots of calls. If you can avoid it (and I know from bitter experience) it’s better not to have your customers perform the stress tests on your behalf. What price peace of mind.

    The Dory call simulator can be setup to initiate multiple calls per second and build scenarios  that replicate real life situations. It empowers you to decide whether to roll back an upgrade before it is customer affecting.

    The other cool feature is Dory’s ability to test CPE such as PBXs. If you provide SIP trunks to your customers then the likelihood is that these customers will use a wide variety of equipments at the end of the SIP trunk. The more the type of kit the bigger the support headache. Most operators don’t have the luxury of being able to constrain their customers to a small number of supported CPE.

    The problem is every  time a vendor upgrades their firmware this kit has to be recertified on your network. This is an expensive and time consuming process. Imaging it it could be automated. Well this is what Dory does.

    You can easily construct scenarios that match all the call types handled by a piece of equipment. Call transfer, hold, hunt groups etc etc. You build up your library of scenarios and every time an interop test needs running you just point Dory at the kit and hey presto, it runs the tests.

    There is a little more to it than this but you get the drift. Dory comes with a console, a scheduler and logger, can simulate both inbound and outbound calls and can be run as multiple instances around your network. I have some spiel/data so if you want to know more feel free to get in touch. Also if you think you would like a demo you can request it from the Netaxis Dory page on their website.


    Engineer peering

    @LONAP sponsors NLNOG meeting in Amsterdam

    At NLNOG points mean prizes

    LONAP were pleased to be one of the sponsors at NLNOG 2016 last Friday. This is the Amsterdam gathering of Dutch based Network Operators. NLNOG has around 400 subscribers to its mailing list. There were 180+ attendees which puts it on a par with the UKNOF meetings – if I recall right UKNOF 35 in Manchester had around 170 in attendance.

    We donated a prize to the very hard network engineering related quiz – a wireless hotspot:

    UBIQUITI NETWORKS UBI-UAP-AC-PRO 24/5Ghz 450/1300Mbps 122m

    It was won by SinnerG BV sys admin Mark Scholten – – for coming fourth in the quiz

    We had two guys there and the third prize was won by our very own Richard Irving. As a sponsor he graciously declined the prize but was very pleased to have come third:)

    You may wonder what LONAP, a London based Internet Exchange Point, were doing sponsoring a network engineering event in Amsterdam. Well a great many of LONAP’s recent new members are overseas organisations wanting access to UK content and eyeball networks. Even as I write we have just announced @AIS_Thailand (Advanced Wireless Network) – the first network to connect to @LONAP from Thailand.

    Things have moved on from the early days where LONAP was a small UK centric IXP serving UK customers. We still are UK centric actually but as internet traffic grows the rest of the world wants in on the act.

    There are great benefits from including Peering in your internet access strategy. Performance is one and cost is another although it is fair to say that a network has to include their own cost of managing peering on top of the relatively low IXP port prices.

    If anyone is interested in chatting about joining LONAP, a not for profit members organisation of which I am honoured to be chair, please do get in touch.

    A ton of internet peering related posts on this blog.

    Engineer peering

    @LONAP reaches for Sky

    Sky has joined Internet Exchange Point LONAP –immediately boosts traffic

    Pleased to announce that UK eyeball network Sky has joined Internet Exchange Point LONAP. Sky, as most of you will know are one of the largest ISPs in the UK and this move is really an endorsement that the London based IXP has come of age.

    LONAP, which was established 19 years ago, has moved on from being seen as just a peering point for smaller networks in the UK. The membership list now includes the vast majority of the world’s giant content providers: Google, Apple, Akamai, Amazon etc. Now with Sky onboard we are seeing the connection of millions of UK broadband subscribers.

    This news comes after LONAP’s recent announcement of their investment in a brand new 100GigE core based on equipment from California networking company Arista. We now have a healthy number of 100Gig connections in the pipeline.

    LONAP is a not for profit membership organisation operated with two and a half full time employees with occasional contributions from members and the board of directors.

    Check out other internet peering posts on

    If you re interested in hearing more about what LONAP can do for your network please do get in touch. I am trefordavies on Facebook and have had the privilege of having served on the LONAP board for three years or so now.

    btw in case you didn’t get it the featured image is of sky. Blue sky. It is summer after all. There’s a bit of tree in there too:)

    Engineer peering

    UK and Ireland Peering Forum in Dublin – with @dantartic

    UK and Ireland Peering Forum success

    Terrific UK and Ireland Peering Forum in Dublin. Mostly attendees from the republic but plenty of people made the trip over from the UK. It was followed by the regular INEX meeting which had a superset of attendees – bit of a pun – they were all v nice – geddit.

    The day was closed by Danny McFadden (@dantartic) who spoke about his two year stint as IT manager of the British Antartic Survey base in Rothera which is on the sticking out bit. An interesting life experience.

    Rothera has recently upgraded its satellite internet connection to 1.5Mbps which costs them £150k pa. Next time you think of changing broadband suppliers and whinge about pricing just think of that.

    It is currently winter in Antartica and the bases are either closed or trimmed down to a skeleton staff of 18 or so persons. It was interesting to get an insight into the difficult nature of the live in wintertime and perpetual darkness. For example there was at the time of his talk, a medical emergency happening at the South Pole base. An American scientist was ill and needed emergency evacuation to hospital.

    This is harder than you think. Two planes had to fly in to Rothera. One then switched to skis and flew on to the pole. This is a ten hour flight in darkness. It was also a very expensive job. One barrel of fuel costs $1,000 by the time it has been transported to Rothera. To get it to the South Pole costs another $9,000. That’s a ten thousand dollar barrel of fuel. And the plane needs 5 of them to get from Rothera to the pole.

    So the return flight for one plane from the edge of Antartica to the pole cost $100,000 in fuel alone. Science is an expensive business. Governments support it because the Antartic treaty comes to an end in 2050 and countries want to maintain their presence as an investment towards the land grab/bun fight that is expected to happen at that time.

    It could spell disaster for Antartica because they are all interested in the mineral rights.

    As a footnote my sister Sue went to Antartica on a cruise a few years ago. Aside from the fantastic wildlife the high point was the 10 thousand year old ice cubes they used in their gin and tonics – hacked from a nearby ice shelf 🙂

    A big thanks to INEX and especially Eileen for such a well organised event. The UK and Ireland Peering Forum is run by LONAP, INEX, LINX and IXLeeds (in no particular order of preference).

    Other peering posts here.

    Engineer peering

    The great wiring challenge

    Cables r us

    poor wiringThe great thing about working with LONAP is that I get to meet interesting people in interesting places. Yesterday my daughter Hannah and I met with Barry O’Donovan (INEX) and Mara Novakovic (LinkedIn) for early door at Slattery’s pub in Dublin. It’s a big rugby pub that is heaving when there is a match on.

    Apart from the Guinness and the considerable choice of beers the one thing that stood out about Slattery’s was the wiring. I challenge anyone to produce a picture with a worse tangle of wires that in the featured image.

    It’s so bad it’s a wonder anything gets fixed when it breaks – how would you find the right cable?

    Over in Dublin with LONAP for the UK and Ireland Peering Forum. It’s been a very good morning – a morning of talks and opportunities to chat with colleagues, members and potential members.

    Other peering posts on this blog. Pic courtesy of Barry using my phone.

    Engineer internet peering

    First @LONAP 100GigE port connected

    Akamai hook up with LONAP 100GigE port

    Excited to tell you that Content Distribution Networks and LONAP member Akamai have connected at Equinix HEX with our first 100GigE port. This is a testament to the great work done by our engineering team Will, Rob and Tom.

    If you aren’t in the business this might not mean that much to you. 100Gigabits per second of connectivity seems difficult to get your brain around. However if you consider that when streaming HD video from the internet you could be using several Mbits per second and that millions of people are also doing this all at the same time then the core of the internet needs to be able to handle a lot of traffic.

    When it comes to shifting internet data around the bigger the pipe the better and 100GigE is the current state of production art. The project at LONAP is still only part way through. We are replacing our old Extreme switches with a brand new LAN using Arista.

    This is a big investment for LONAP but one that has to be repeated every few years as internet traffic grows and capacity requirements increase. This time around we have been under some pressure from our content provider members to get the 100Gig ports installed. The timing on this occasion is good as the Euro2016 football tournament gets into full sway and folk start watching the matches on the internet. The 100Gig port adds instant capacity. We like to have plenty of headroom on our network at LOANP.

    The member and port count at LONAP continues to grow up and to the right. It is a good place to be right now. We are very lucky to have a great community of members. If you haven’t yet registered but plan on coming to the UK and Ireland Peering Forum in Dublin next Monday I’d get your name down. See you there.

    PS For what it’s worth I’ll be helping to reduce the pressure on the internet by watching the Wales v England game on the TV in the beer garden at the Strugglers pub in Lincoln. C’mon Wales.

    More peering posts on

    Engineer peering

    UK & Ireland Peering Forum – Monday 20th June

    A chance to talk peering in Dublin

    Peering in Dublin. If you haven’t already got your name down and work for an organisation that is a member of one of the UK Internet Exchange Points you should seriously consider coming along to the UK and Ireland Peering Forum. This is being held in Dublin on Monday 20th April.

    Members and prospective members of INEX, IX Leeds, LINX and LONAP will be gathering for a day of talks and discussions on subjects of real interest to the network engineer.

    The UK&I Peering Forum takes place in the morning and participants are invite to the INEX meeting in the afternoon. Regular participant at UKNOF meetings will be familiar with some of the regulars who come across for the events. This is a chance to meet and talk with the rest of the Irish networking community.

    We are going to be covering peering tools, take a peek at what is happening on the regulatory front and how this might affect they way you have to run your network, together with a couple of technical sessions.

    The biggest benefit you will get from this meeting is the opportunity to set up new peering relationships. Attending will also allow you to meet your peers (pun intended) at other networks and catch up on what are the issues of the day.

    I will be announcing the full peering in Dublin line up in the next few days.

    Peering in Dublin – UK & Ireland Peering Forum Hosts

    There are four host exchanges of the UK & Ireland Peering Forum. These are INEX in Dublin, Ireland, IXLeeds in Yorkshire and LINX and LONAP who are both based in London.


    INEX is a neutral, industry-owned Association, founded in 1996, that provides IP peering facilities for its members. INEX membership is open to all organisations that can benefit from peering their IP traffic. There are currently 96 members. Our objective is to provide high-speed, reliable and resilient IP traffic exchange facilities for both Irish and International organisations, allowing them to route IP traffic efficiently thereby providing faster, more reliable and lower-latency internet access for their customers. The INEX switching centres are located in six secure data centres around Dublin; Telecity Group in Kilcarbery Park, Citywest Business Campus and Northwest Business Park, Interxion DUB1 and Interxion DUB2 in Park West, and the Vodafone data centre in Clonshaugh. The infrastructure is connected by dedicated resilient fibre links.

    For more information, please visit


    IXLeeds is one of three internet exchange operators in the UK and the only operator based outside of London. It’s aims are to promote regional cooperation between network operators and to raise awareness of the purpose and importance of a solid internet exchange fabric in the UK. IXLeeds is a limited company with a board of four directors and a company secretary. IXLeeds’ first board was elected on 21st September 2011 by its founding members and comprises of Andy Davidson, Allegro Networks (Chair), Mark Fordyce, York Data Services (Finance Director), Tom Bird, Portfast (Director), Thomas Mangin, Exa Networks (Technical Director) and Adam Beaumont, aql (Company Secretary)..

    For more information, please visit


    The London Internet Exchange (LINX) is a global leader of Internet Exchange Points (IXP). Our company ethos and aspirations result in constant growth and improvement in our services and as a result, members have access to the most attractive peers at the lowest cost. With over 690 members connecting from over 66 different countries worldwide, LINX members have access to direct routes from a large number of diverse international peering partners. In addition to its dual LAN topology in London using equipment from Juniper Networks and Extreme Networks, LINX has three UK regional exchanges (IXManchester, IXScotland and IXCardiff) and another in North Virginia, USA (LINX NoVA). Using LINX allows a reliable exchange of traffic with increased routing control and improved performance. As a not-for-profit organisation, we focus on investing our service and membership fees into strengthening LINX network services. This ensures that our infrastructure is as up to date as possible and that we remain at the forefront of the IXP industry. By doing so, we can confidently provide our members with improved network performance, low latency and more control.

    For more information, please visit


    The London Access Point [LONAP] was first established in 1997 as a ‘not for profit’ Internet Exchange Point for London. Today, our list of active members includes global brands, London businesses and FTSE100 companies, all of which are joint stakeholders in the organisation. As members, they all have exclusive interconnectivity and direct input into the configuration of the network and enjoy the commercial, operational and social benefits of working with a vibrant and dynamic exchange. Our growing membership includes ISPs, network operators and content providers with their own data networks. LONAP members exchange traffic using a network of interconnected switches hosted in our data centres across the City of London and Docklands.

    For more information, please visit

    Engineer fun stuff peering

    LONAP @ RIPE72

    LONAP is at RIPE72

    Wearing my LONAP hat again today. Actually that is a figure of speech. You can see from the featured image that I’m really wearing my LONAP shirt. Facebook friends will know that I’m at the RIPE72 conference in Copenhagen.

    We, LONAP, get a lot out of these conferences. Not only is there a lot of good content but it is a fantastic place to meet existing and prospective members. The LONAP community is growing.

    LONAP are also slap bang in the middle of a total core network upgrade. Our new network is going to be based on Arista 100GigE kit. 100GigE has been around for perhaps four years but the first generation of equipment was very expensive. The introduction of cheaper more powerful silicon has brought the pricing down with the timing being just right for our roll out. It feels as if 100Gig is only now reaching the same stage of commoditisation that 10Gig was at 7 years ago.

    Other benefits that Arista bring include programmability (API) and VXLAN for loop free layer-two. It’s worth noting that the Internet Exchange Point market has specific technical needs that aren’t addressed by all vendors.  The fact that Arista has an industry category specifically for identifying IXPs in its customer sign up page is very telling. I’ve not seen this in any other vendor (am prepared to be corrected here).

    Check out the image


    It’s an exciting time to be at LONAP. Our IXP in the middle of a transition from being “just a small exchange run by people with other day jobs” to a professional outfit that is attracting big players from both the content provider and eyeball network communities.

    We still like to think of it as a family business though. We are a community that does things on behalf of the community. Our low overhead base means we are amongst the most cost effective IXPs in the game.

    More in due course. In the meantime if you are  at RIPE72 and want to chat to us look out for our LONAP branded shirts.

    Check out all peering posts here.

    ecommerce Engineer internet ipv6

    NANOG 66 is in San Diego by the sea @LONAP

    NANOG 66 – bring your shorts

    One of the downsides perhaps of being in the internet plumbing industry is that your Facebook timeline constantly fills up with people  off to conferences in exotic places. Even when they aren’t off to a meeting people are still off to exotic places as they spend their air miles. This week it’s NANOG 66  in San Diego.

    San Diego is one of the nicest places you can choose to go to a conference. The other popular destination is Hawaii. I’ve done San Diego but not Hawaii. In fact I tend to avoid conferences that involve long haul travel these days because it’s hard enough submitting your body to a week of conference talks, corridor meetings and the follow on eating and drinking without throwing jet lag into the mix.

    Out of curiosity I took a look at the NANOG Facebook group and the NANOG website. The first thing that hit me on Facebook was a comment that told me there were 1002 people signed up for the conference. That’s a lot of people. So many people that you have to figure out how to make the most of the fact that they are all there. The value in these meetings tends not to be in the actual conference material but in the people that attend.

    This brings me to my second observation. The list of Platinum Sponsors, organisations that have paid $100,000 to sponsor the three events in the year, includes a business called Addrex. Addrex are an IP address broker. That’s a serious chunk of change they have invested in marketing by becoming a NANOG sponsor. It just goes to show how much money there must be in the IPv4 address game.

    internet traffic growthI am more familiar with Sandra Brown and the IPv4 Market Group who very generously sponsored the LONAP dinner last month. LONAP usually sends a couple of people to the NANOG meetings because they are very good recruiting grounds for new members. Most of our big members have come on board as a result of relationships developed at such meetings.

    If you are in NANOG this week do take the opportunity to look up Will Hargrave and Richard Irving who will be there representing us.

    Engineer peering

    LONAP networking dinner is great success @JoeBaguley @LONAP

    Industry peers get together for LONAP networking dinner

    The LONAP networking dinner held in a private dining room  at Kettner’s last Thursday night was as usual a huge success with guest speaker VMWare CTO Joe Baguley providing some stimulating thoughts for debate. Joe was speaking about Network Function Virtualisation.

    LONAP as you may know is an Internet Exchange Point (IXP) of which I have the honour of being Chairman. Now the whole point of an IXP is to provide eyeball networks and content providers a meeting place for them to share their network traffic (peer) cost effectively and get the best performance out of their networks. The typically lower latencies obtained by peering at an IXP result in the best experience for their end user customers.

    Joe Baguley in his talk highlighted an interesting use for NFV that was highly analogous to the way an IXP works. He discussed the scenario where a betting firm could set up a virtual network at the datacentre closest to a specific sporting event. This would in theory give the betting firm’s customers the fastest response time and best experience. ie help them to lose their money more quickly.

    Once the event is over they just tear down the virtual network. A click of a virtual switch. Highly cost effective.

    Virtual network functions we are told are just as good as hardware implementations these days. When you think about it unless your need requires high end state of the art ASIC silicon everything is realisable in software these days.

    These LONAP networking dinners are always successful, helped on this occasion by the very generous support of the IPv4 Market Group (IPv4 address proker) and Xantaro (network integrator). I think both sponsors will have had great value from the exposure into a fast growing and high spending community of network operators. Take a look at their sites.

    The attendees were a great mix of LONAP members and non-members/prospects ranging from small but agile communications providers to some of the biggest eyeball networks and content providers in the country.

    These are top class networking events. If you are in this business, a LONAP member or otherwise, look out for the next dinner which will be sometime in the spring. If you want to sponsor a LONAP networking dinner by all means drop me a line. Vendors get really good exposure into a wide network engineering community.

    Last week’s dinner did have a hint of that end of an era feeling. It was the last  week  of trading for Kettner’s, a legendary Soho bar that has been around for 149 years. Kettner’s has been sold to the SoHo House Group who are turning the whole block into a hotel. This is a shame because I’ve been holding private dinners for a few years now and only discovered Kettners in the last year. I’ll have to find another venue.

    The sadness at the closure of Kettners was tempered by the fact that because they had been running down stocks of wine we got the most expensive plonk on the wine list for the lower price of the wine we had ordered but which had run out:)

    Ciao amigos. Keep peering! A few pics of the evening below including some colourful ones taken en route. Oh and a big thank you for Joe Baguley for coming along and speaking:) Loads of peering content on this blog if you want to read it.

    LONAP networking dinner

    LONAP networking dinner
    LONAP networking dinner guest Joe Baguley gets animated

    LONAP networking dinner

    LONAP networking dinner

    LONAP networking dinner

    Engineer engineering internet

    UKNOF33 – live action from the conference room floor

    UKNOF33 is happening today at Bishopsgate deep in the heart of the City of London

    UKNOF33 returns once more to Bishopsgate. It feels a bit odd for what are obviously great bunch of guys and gals, famed for their altruism and generous heartedness to be holding a conference deep in the heart of the financial centre of London – a place known for its single minded pursuit of the filthy lucre purely for the sake of it and (allegedly) at the expense of all moral consciousness. That isn’t to say your average network engineer isn’t interested in the green stuff but we do like to think that we perform a good public service as part of the deal.

    Nothwithstanding the location, which due to the growth of the conference is one of the few places able to hold all the people wanting to come, UKNOF continues to be one ofthe most worthwhile meetings in the networking calendar.

    These UKNOF posts, in line with other conference posts, aren’t necessarily going to cover the presentations verbatim. You can download the slides if that’s what interests you. No, in these posts you get whatever catches my eye. This might be corridor gossip, interesting (editor’s judgement is final as to what constitute interesting) snippets from the talks, or just photos of things that catch my eye.

    Feel free to share, comment and generally participate, if you want to. Otherwise it’s good to be back 🙂

    If yer interested check out our other UKNOF posts.

    Engineer internet

    RIPE71 social scenes

    It isn’t all work you know – videos from the first night RIPE71 social

    The RIPE71 social was a goodun. It’s always a bit of a risk using video recorded at such events because you never know what clarity of diction is going to be achieve, if you get my drift. On this occasion we were all very sensible and I think hte vids are ok to go:)

    Engineer internet travel

    RIPE71 – end to end reportingish

    A RIPE71 story begins. The live blogging feature doesn’t seem to want to show embedded videos so these are pasted into the body of the blog

    Engineer voip

    Nominet kills off ENUM project

    Nominet ENUM is no more um

    This arrived in the email this morning from Nominet Head of Policy David Abrahams via ITSPA:

    “As you will be aware the Nominet ENUM project of 2007 was intended to provide a method for end-users to register ENUM domain names in the zone. Nominet signed a contract with the UK ENUM Consortium in 2007 to operate the infrastructure for the UK ENUM system for an initial five year period.  Although the UK ENUM Consortium went into liquidation in 2012 we have continued to run a provisioning system (EPP) and dynamic updates and nameservers to support ENUM. 

    The registration process for ENUM requires a participating registrar and a Validation Agency.  The Validation Agency has to be appointed by the UK ENUM Consortium.  There are currently no ENUM registrars or Validation Agencies operating for UK ENUM and there have only ever been 6 external registrations since the service was launched, that last being in August 2011.

    We therefore intend to suspend support for ENUM services in November, to reduce our overheads of maintaining the infrastructure and service at a time when there is no prospect of any further registrations in the near future.   A copy of the DNS zone as it stands now will be held in a repository in the event there is a decision or sufficient demand in the future to resurrect this service.  We will endeavour to contact the registrants to notify them of the suspension.

    We have made DCMS and Ofcom aware of our intention and thought we would let you know for good measure.”

    This is kind of sad. Those of us who were around in the early days of VoIP and SIP had big hopes for ENUM, the system whereby voice calls would be routed over the internet.

    The biggest problem was in those days I worked for a PBX vendor, Mitel. PBX vendors have no vested interest in which medium actually carries a phone call or whether that phone call is free or not. The ENUM proposition was that calls would simply be data sessions and therefore free over the internet.

    This didn’t really fit in with telcos/carriers who make their money, and in fairness recover their overheads, by charging for calls. It’s no wonder that ENUM didn’t make it as an universal free service. That doesn’t mean to say that ENUM-like tech is not used by carriers to run their voice services as part of their core network.

    It is also clear that free Over The Top voice services can be successful – witness Skype, Google Hangouts , Apple Facetime etc etc etc. It’s just that these services have been better marketed by organisations with very deep pockets and other ways of generating their cash.

    Still I had a quiet moment of reflection in advance of writing this post in memory of Nominet ENUM and the people involved in the project.

    PS featured image is an old stock photo mentioning Nominet. The one of me and Vint Cerf at a Nominet conference was too small to use. Check out the post I wrote following the Nominet ENUM project launch.

    Engineer peering travel

    Lovely to have you with us

    Return from Euro-IX Berlin

    I don’t know whether it’s just hotel living or the exposure to different cultures but it’s always nice to come home after a trip overseas. For the first part of this week I was at Euro-IX Berlin. It was the usual useful meeting of Internet Exchange Point operators from mostly Europe but also around the world and I always get a lot out of it.

    Our hotel in Berlin put on a fabulous spread for breakfast every day and indeed I took along a bottle of HP Sauce knowing that outside the UK you rarely find such condiments at the table. I knew however in my heart of hearts that the bacon and sausage on offer that would go with the HP were not going to meet spec. They didn’t. After the first day I stuck with croissants and fruit. Very sensible really.

    I also find myself making up sandwiches for lunch. These conference hotels always put on a substantial buffet but just think about it. How many of us have a large lunch at work every day (ok outside of France, Spain, Italy etc 🙂 ). We usually have a sandwich or similarly light nosh. So in Berlin I took some bread and whatever meat was on offer and just made a butty.

    Dinners I am ok with. Usually international fare served up in every country and accompanied by moderate amounts of beer and wine in keeping with being compos-mentis for the next day of conferencing (bed by 11pm etc :)). However how many of you go out for dinner or to the pub on several consecutive nights when you are at home? Right.

    So whilst we all enjoy a bit of travel we all like to get home to our own comfort zone. In very recent memory I recall coming back from consecutive trips to Bucharest and Toledo. After having had a week of rich living I was beginning to suffer (I’ll leave it to your imaginations to decide how). My first action back in the UK was upon leaving Luton Airport to stop at the services on the M1 and purchase a cheese sandwich and a pint of milk. After consuming said items my constitution returned to normal. Hey presto. Amazing really the effect of a cheese sandwich and a pint of milk.

    On the trip to Berlin the recuperation began the moment I entered the BA lounge at Telgel Airport. Then when on board the captain welcomed us over the tannoy saying it was “lovely to have us aboard” my shoulders began to relax.

    I’m not really a big fan of British culture especially when abroad but there was something very comforting in that use of language. You aren’t going to get it anywhere else. It was as if Sgt Wilson from Dad’s Army was at the controls. Lovely.

    The trip home was greatly assisted by technology – Uber. It’s got to a point when travelling that its a real disappointment to arrive somewhere that doesn’t have an Uber service. The rest of the day was beans on toast, hot bath and early to bed to avoid the Apprentice on TV (which really annoys me).


    PS check out my talk at Euro-IX Berlin about getting stuck in a lift.

    PPS the featured image is of some grappa being poured for me at the Aigner restaurant on Tuesday night

    Engineer engineering internet

    UKNOF32 Sheffield Day 2

    UKNOF32 Sheffield Day 2

    Engineer peering

    UK & Ireland Peering Forum Sheffield

    Live blogging from the UK & Ireland Peering Forum


    Engineer Mobile mobile connectivity UC

    Mobile Unified Communications Network Architecture – an Operator Perspective

    Introduction to Mobile Unified Communications Network Architecture

    Vodafone Group Solutions Architect Juan Hernandez presents us with an authoritative look at mobile Unified Communications Network Architecture

    Unified communications (UC) is a hot topic in the industry. Although the provided services may be pretty much the same in different market niches, the way in which such services are implemented can vary significantly.

    This post is aimed at providing the perspective of a mobile operator. We cannot talk about UC in mobile operators without talking about IMS (IP Multimedia Subsystem). IMS has become a standard across the mobile operators for a number of reasons. Although IMS was designed to integrate SIP devices and obviously 2/3 G mobile phones are not SIP end points, the arrival of the 4G standard (LTE) has introduced a new and interesting actor in the scene, the VoLTE device. VoLTE (Voice over LTE) implementation is the most important challenge the mobile operators are facing nowadays. This type of device is a SIP device, what brings mobile phones into the sphere of native end-points manageable by an IMS.

    This post does not intend to be a comprehensive description of the UC implementation with IMS, but an introduction to the IMS as a concept and to the role it is playing in UC in the context of mobile operators.


    IP Multimedia Subsystem

    This name highlights two important aspects:

      • IP → IMS is based on end to end IP connections.
      • Multimedia → Different media can be combined, like voice, video, images, text, etc.

    It means that “circuit switch” (CS) technology has been overcome and that there is not a split between voice calls and data connections any more. All media formats are managed by means of IP connections instead.

    Separate and standardized access, routing and service planes

    In my opinion, this is the major advantage we get with the introduction of an IMS. The three functions get separated and independent: access, routing and service. Standardized modularity allows  the combination of different vendors and access technologies,  and makes network evolution more flexible.

    Mobile Unified Communications Network Architecture separation of service routing and access planes

    Fig.1 Separation of access, routing and service planes

    Multi-access and multi-terminal

    IMS platform is access independent. We will see later on how the IMS elements P-CSCF and AG (access gateway) play the role of proxy, to face SIP and other VOIP protocol devices (H323, mgcp) respectively.

    Mobile Unified Communications Network Architecture multi device access



    Fig.2 Multi-access and multi-device within an IMS

    The previous figure shows how the P-CSCF interfaces all SIP connections, independently from the type of connection or device in use. Embedded SIP clients can be used in 3G networks. In this case a data session is established between the mobile network and the P-CSCF. On the other hand LTE devices are able to establish SIP voice calls directly with the P-CSCF. It is important to highlight that VoLTE devices can place SIP calls in a native way, without the need of embedded SIP clients.

    SIP Internet connections can be also established against the P-CSCF. So, voice calls (VoLTE) and voice calls encapsulated into data connections can be managed by the IMS.

    On the other hand, the AGCF (Access Gateway Control Function) can manage other VOIP protocols, translating them into SIP in the boundary of the IMS space.

    Of course, SIP phones can also be managed by the IMS by means of the P-CSCF.


      • Of services
      • Of billing

    Mobile Unified Communications Network Architecture service convergence


    Fig.3 Service convergence

    The previous figure shows the way in which both, fixed and mobile devices, can use exactly the same services, provided by the AS (Application Server) in the service plane. Therefore, the user experience is the same, independently from the user device.

    Mobile Unified Communications Network Architecture billing convergence


    Fig.4 Billing convergence

    It comes evident from the figure 4 that once the same Application Server provides the services to fixed and mobile devices, the CG (charging gateway) can get convergent billing data for fixed and mobile networks.


    IMS architecture is anything but simple. The already commented characteristics of function separation and modularity leads to a complex architecture that can be seen in the next two figures.

    Mobile Unified Communications Network Architecture IMS functional modules


    Fig.5 Functional modules in IMS


    Mobile Unified Communications Network Architecture IMS

    Fig.6 IMS architecture


    Next, I proceed to explain the roles of the most important elements in an IMS:

    P-CSCF (Proxy – Call Session Control Function)

      • Access contact for SIP native clients
      • Security control
      • QoS Policy function.

    The IMS space is considered to be a trusted area, since the P-CSCF applies a series of security procedures towards the access plane.

    The QoS policy function is applied when a Policy server is deployed, in order to apply QoS policies in the access network. The P-CSCF communicates with the Policy server to orchestrate the QoS policies.

    I-CSCF (Interrogating – Call Session Control Function)

      • Allocation of the S-CSCF to serve the user trying to register
      • Allocation of the S-CSCF serving an IMS called user

    I-CSCF accepts requests for registration and terminating calls, and it interrogates the HSS (Home Subscriber Server) to retrieve the corresponding S-CSCF address.

    S-CSCF (Serving- Call Session Control Function)

      • User registration
      • Authorization and authentication of users
      • Service triggering.

    The S-CSCF is in charge of registering a user and triggering the services for such user by routing the incoming requests to the Application Server.

    HSS (Home Subscriber Server)

      • Service profile
      • Subscription data
      • Application Server allocation
        • HSS registers the S-CSCF assigned to a user during the REGISTRATION procedure
      • Authentication information
        • Authentication tuples calculation and delivery of such authentication tuples to the S-CSCF

    The HSS is the data base where all the information related to a subscriber is provisioned. It is the equivalent to the HLR for mobile networks.

    MGCF (Media Gateway Control Function)

      • Interworking with PSTN (signalling)
      • MGW control by means of H.248

    The MGCF is in charge of interworking with PSTN by exchanging ISUP messages. It is also in charge of managing the allocation of ports in the Media GW to support the media interworking between the IMS and PSTN. This control of the Media GW is based on H248 protocol.

    MGW (Media Gateway)

      • Interworking with PSTN (bearer)

    The media gateway offers TDM connectivity against PSTN and IP connectivity for RTP flows with the IMS devices.

    AS (Application Server)

      • It applies the supplementary services
      • It applies Unified Communication services

    So, in the context of IMS, while the UC services are applied by the Application Server, the routing and access management functions are done by the IMS.


    Unified communication services can be implemented in a number of ways. The major distinction can be done depending on whether the service is deployed locally within the end user premises, or it is hosted in a central platform serving a lot of local customer offices. In the first model a series of local servers (presence, telephony, instant messaging, etc.) are deployed in order to provide the service to a single office or to a series of premises within the same company. In the second model, an Operator company hosts the service for a big number of customer sites. The latter is an Operator model that can be deployed in several ways.

    The traditional way to implement UC services by an Operator company is by using a compact softswitch in which routing, provisioning and service planes are put together into a compact platform. CS2K by Gendband and Broadsoft Application Server (stand-alone mode) would match with that model.

    As we have seen in the previous sections, IMS is the new way in which mobile Operators are deploying their services nowadays. In this model, access, routing and service planes have been separated. Now, the service is only associated with the application server placed in the service layer. The previous cited vendors, Gendband and Broadsoft, have IMS compliant implementations. In both cases the service engine (application server) is used connected directly to an IMS. So, routing and other capabilities in the stand-alone product are not use in this implementation. Of course, a series of requirements like offering a standard isc interface to the S-CSCF are required in an Application server to be compatible with any IMS vendor.

    As summary, I would say that the understanding of how mobile operators are deploying UC services nowadays, requires the understanding of IMS technology, what is not an easy challenge. IMS modularity brings flexibility but it leads to a higher level of complexity in the way in which the different modules interact among them.

    Juan Hernández (Solution Architect at Vodafone Group)


    Twitter: @unveilingthereality