© 2022 Black Swan Telecom Journal | • | protecting and growing a robust communications business | • a service of |
Why is B/OSS software so interesting to follow? Because it’s an endless game of leapfrog.
Software that was state-of-the-art only five years ago is rendered obsolete by new networks, fresh customer needs, or a hundred other market shifts. These shifts often create a habitat for new software solutions and players to live.
Edoardo Rizzi, Vice President of Trendium headquartered in Colorado, believes mobile’s shift to LTE is a golden opportunity for innovative software to hatch and swim about profitably. Edoardo knows much about the market he’s pursuing being a fifteen-year veteran in the industry and former business manager at Tektronix Communications. At Tek he was on a team that integrated Tek’s older generation systems with INET, the star product that enabled Tek to eventually lead the 3G service assurance market.
Edoardo has taken on a bit of a career challenge coming to Trendium, a company that relaunched itself at this year’s MWC in Barcelona, because he believed in the company and in Sameh Yamany’s vision (also a former Tek and industry veteran), and what they are trying to do for the future of operators. The system he discusses in this interview is in trial phases, though Trendium is no stranger to operators like AT&T who have used its Business SLA Management system for more than ten years now.
I think you’ll enjoy Edoardo’s perspective: he not only clearly explains how LTE is disrupting the network monitoring scene, but also gives us lessons on how an OSS software firm strategizes and carves out a niche for itself.
Dan Baker: Edoardo, what’s so different about LTE networks that call for a paradigm shift in service management systems? |
Edoardo Rizzi: Dan, LTE and the emerging heterogeneous networks are surely going to shake up the service management and CE (customer experience) assurance worlds. “Elastic” is perhaps the best word to describe those future networks: almost instantaneous decisions will be made to provision capacity, limit what a user can do and consume, or decide how much power a base station will transmit.
In 2G and to a certain extent in 3G networks, you could afford to over-provision for voice services because it was easy and it didn‘t cost you very much. But can you really over-provision for tomorrow’s mobile broadband? You can try, but you’ll find that your profit margins will totally disappear.
There are two reasons: 1) it’s hard to predict which pockets of an LTE network will see spikes in demand; and 2) LTE revolves less around voice and more around data, for instance streaming video — which eat up 100 times more bandwidth.
How will operators manage these unpredictable and bandwidth-hungry traffic shifts? |
The solution is to deploy capacity in real-time where it’s needed. Capacity planning is vital, but it only takes you so far. When a special city event occurs like a marathon or large parade, how do you deal with the demands those events put on your network? An operator would need to dynamically manage the macro-cells and small cells across the region to deliver the best experience for people attending the special event, and since spectrum is scarce, you may need to dynamically juggle that resource too.
What we’re talking about here is having “real-time awareness” in areas of the network where you want to make real-time decisions. This is one of the most demanding “big data” applications out there because the volumes are huge and decisions need to be made in a matter of seconds.
A monitoring and correlation engine lives with and enables many other in-network applications. Many network entities need real-time visibility, so it’s the management system’s job to feed those systems, whether they are for clearing network congestion or enabling self-optimized systems.
Operators also must focus on providing assurance. If I have entities that make network configuration decisions in real-time, they are going to affect customers. This means I need to keep a pulse on the real-time experience of certain affected customers in real-time. Consequently, finding out 30 minutes later that somebody was not happy is not acceptable. You need to take action to mitigate any adverse affect from all the network events and real-time decisions you’re making.
It’s also no longer adequate to monitor things at the core of the network alone. Increasingly, the action is going to happen in the RAN (Radio Access Network). When you try to push traditional monitoring tools into the RAN, they don‘t scale — financially, and often times technically because those monitoring technologies were conceived at a time when the network was mainly carrying voice.
What’s the problem with the current generation of network monitoring systems when it comes to LTE? |
Perhaps the biggest limitation is the speed at which you can analyze the data.
Here we believe a paradigm shift is needed in terms of the way traffic is analyzed instead of doing it based on CDRs (call detail records).
When voice was the major service, CDRs were the way to track, and the previous generations of service management systems are all built around the CDR concept. Trouble is, when you push a CDR-analytic framework into an LTE network that is mostly IP data, you lose efficiency — it costs you a lot in memory and computation time and doesn‘t provide you any added benefit.
For every message that comes in with CDR tracking, you need to ask, “Which session does this message belong to?” Then you must find all the open messages and put it in the right bucket. Plus the sessions are longer now because these are web data sessions now, not just voice, meaning the buckets need to be enlarged.
The bottom line is, you need a bigger, more expensive probe to maintain the CDR way of tracking, and CSPs expect much more from LTE and future mobile broadband networks. In fact, the expectation is that -- on a per unit of bandwidth basis -- the costs of the network and the associated monitoring and traffic analysis systems will decrease by a factor of ten!
So how do you get better price/performance in your approach? |
It begins with our Deep Packet Inspection (DPI) probe: we are getting an order of magnitude speed/cost advantage from it because we use our own, patented, line-rate packet analysis and traffic analysis technology. Also, we have architected our solution so that we can leverage the latest off-the-shelf processing and computing technology available out there.
Having our own packet analysis and traffic analysis technology not only gives us better control of the traffic as it comes in, it also greatly simplifies downstream processing. For example, we can process 100 Gigabit per second in a single rack unit of footprint, and we don‘t need to sample: we already have immediate access to all of that data.
Finally, our architecture delivers greater price/performance benefits to our customers because it was build ready to be virtualized in its entirety, including the traffic analysis agents.
What sort of product experience is Trendium taking into this fresh LTE market? |
Dan, our mediation and correlation engine is something we’ve developed over the past decade, and it’s a true real-time platform: it doesn‘t use database technology to calculate KPIs and policies. Everything is done in memory, which means our correlation and mediation systems can analyze the data in real-time and produce results and insights faster.
Today, AT&T is using that platform for SLA management and service performance management. When it was first sold, it looked at Frame Relay/ATM, then later moved to IP and optical. More recently AT&T has deployed it to monitor SLAs for its enterprises in the Metro Ethernet space.
What we are focusing on and striving for is to take that real-time experience into the mobile network domain.
Aside from better price/performance in analyzing the traffic, what other capabilities do you consider a competitive advantage of the previous generation of network monitoring and expert systems? |
One factor where we certainly plan to excel is providing a deeper level of guidance to the users as they analyze the data whether it is for assurance or analytics purposes. Current service assurance dashboards are great, but for LTE, the amount of data is simply too large and the operator needs more help isolating the biggest problem areas and then quickly finding the root cause of problems.
Our application provides first a view from the outside in. The question there is “What’s the impact from a subscriber’s perspective?” Here we can help the user see the situation through the eyes of the subscriber and identify the biggest problems that are worth investigating.
The next question is “Where are the problems occurring?” Is it the RAN, the core, or the device? This fault domain isolation step is very critical because nowadays phones and tablets are very powerful devices and can themselves be the source of a problem.
Also, in almost any interaction the subscribers have with the network, they are interacting with the back end of an application domain that usually resides outside the scope and responsibility of the network provider. So for the network provider, knowing who’s responsible for a problem has become critical so you don‘t waste your internal resources. Once you identify the domain, you can quickly drill down to the root cause, and that’s exactly what we help them do with our tool.
Another advantage operators find useful is our openness. From the beginning we assume a heterogeneous management system. Different probes, different management system, different elements; we can collect from all of them and we can feed data to any application or systems that needs it.
Many decisions are being driven by forces outside our control — such as marketing or capacity planning — and we know we can‘t influence those, however we do have APIs that allow carriers to feed any application they want to us.
Just about any large carrier has probably 20 to 30 systems they use to keep an eye on service performance and network assurance, so we need to fit into that environment — fit into whatever they have currently deployed, and offer them a path to some ideal future state.
Maybe a carrier deployed a certain vendor’s probes for the 3G network. Well, for LTE, Trendium would deploy its own probes and then take feeds from the existing 3G system in order to provide the service provider with an integrate view across 3G and LTE and at the same time allow the operator to maintain his 3G investment.
We can also take data directly from the network elements. In the RAN we can have probes deployed at the base stations aggregation sites that can monitor wireless traffic backhaul. We also have visibility to what the eNodeB’s are doing. That’s key because a lot is done there without letting the core network know. That’s how LTE works.
Then there are the agents on the devices that capture user experience, network performance, and device health data. So correlating all these data sources gets us an integrated view from an end-to-end perspective.
What’s to prevent Tektronix, NetScout and others from adopting your approach and beating you in the market? |
There’s always a risk that rivals will beat us to market. For many years I worked for a large company, so I have a good idea how large, established software companies operate. In general, existing players will not react quickly enough. It requires them to totally re-architect their product, but a large successful company doesn‘t want to do a radical re-write of an application that’s a market winner.
A large vendor’s success also presents a problem for them. The large installed base of customers tends to drive a lot of what they do and don‘t do, and if it’s a choice between developing new, innovative software for the future vs. developing additional features for a customer who is dangling $10 million in front of you, the short-term money becomes very attractive.
This is the reality and those are some of the market dynamics. We also recognize that when you innovate, not everybody is willing to jump aboard right away. They need to see a few big guys embrace our technology.
We talk to leading telcos in North American and Europe who have experience with LTE. We know how much pain the Verizons and AT&Ts of this world are feeling as they try to manage these immensely complex LTE networks in real-time. Trendium has the solution to this painful transition.
Edoardo, thanks for your honest insights. I wish you — and your competitors — the best of luck. May the best solutions emerge victorious. |
Copyright 2013 Black Swan Telecom Journal