I’ve made the argument twice now that Oracle’s acquisitions of Acme Packet and Tekelec are Big Data plays. I stand by that argument, but I also think these buys are a sign of how intelligence is moving out of the network and into a sort of new layer that pulls together things like diameter signalling and session border control, policy control, software-defined networking, and real-time billing and charging.
BillingViews reader Geoff Coleman made the point, in his recent comment, that a variety of large players from the IT side of the business are “moving the policy (management) away from the network side of the house and more to the business side.” I agree with his observation. The reason I think this trend is emerging, however, is because of this background trend of operators essentially trying to get out of the network infrastructure business.
The dirty little secret that has really bit operators in the backside in the past few years is that it is awfully expensive to build and operate 4G/LTE networks. It is even more difficult to amortize investments in new networks because technology is advancing so rapidly. During a recent chat, Clarity’s Tony Kalcina made the astute observation that operators are also frustrated by regulation. They’ve spent billions and taken on all of the risk to build these new, complex networks but aren’t allowed to charge the OTT guys what they feel is a fair share to help pay for them. In their eyes, the OTT guys are getting a free, or massively subsidized ride, while the network owners lay out the cash, eat all of the risk, and do the hardest work. One of the ways we see operators responding to this conundrum is to push operation and maintenance of their physical networks over to their major vendors, like Alcatel Lucent, in long term managed services deals. Kalcina points out that we also see initiatives across geographies where governments are investing in shared network infrastructure and where operators are beginning to share network resources to gain greater economies scale.
In the meantime, conventional wisdom says that network hardware is becoming more of a commodity, especially as we shift to all-packet networks. That’s part of the reason why we see big NEMs, like NEC and Ericsson, buying OSS/BSS suppliers to move up the stack and provide more of an end-to-end and value-added offering. They know that the business is now more about what you do with logical networks than physical. Back in the day, there wasn’t much of a separation between the two. But the more we adopt packet technologies, virtualization, multi-tenant IT platforms, shared data centers, public clouds, and so forth, the more logical networking – and logical networking that can be set up, configured to spec, and torn down in real-time – becomes the real point of differentiation.
Now we see this ‘new layer’ – if we dare to call it that – coming into focus. In that new layer lives stuff that inhabits a space that’s perhaps between the network and what we usually think of as OSS/BSS. Maybe that new layer includes all of the things that devices like diameter signalling controllers and session border controllers do, along with policy managers, software defined networking platforms, and real-time transaction-oriented gear – like charging systems, and digital rights managers. There’s likely a bunch of other smart, useful technology in there as well that either originated in the network and moved up, or in the IT stack and moved down (closer to the network).
So, going back to Oracle…When it bought Acme Packet, Tekelec and don’t forget Xsigo (and who knows which company next) it staked its claim on this new layer. In a sense, it’s buying up pieces of the new network business without actually getting into the network hardware business.
I still think that these are Big Data plays, because Oracle is a horizontal company and at a strategic level this is all about who controls the cloud-computing space (which is fast becoming a hybrid of data center and network technology) and thus who controls access to information and information about how people, devices, and companies are using information. But within the telecom vertical, Oracle’s acquisitions make a heck of a lot of sense given the global market now seems to value intelligent, software-based, logical networking over physical networks.
Ed,
I think you are completely on the money with regard to both Oracle’s Big Data play, and especially the ability it gives to monitor, analyse and, most importantly, control and monetise the transactions and data flowing across the networks, and to the reasons why the NEMS are moving in the same direction.
However, I would counsel caution about how soon this brave new world will arrive; integrating the different products and tools together so that they successfully implement this vision will take time, as shown in nearly all of the past software product acquisitions by the OSS/BSS vendors and the NEMS.
Bill Mumford
Agreed Bill. Strategy is one thing. Execution? Quite another. And what happens when you assimilate two solid brands into a big brand not necessarily associated with the expertise for which the two smaller brands are best known? Lots of questions there.
Hi Ed,
It’s very interesting your point of view about this new layer where the charging, controlling, network, etc information are gathered and treated. Big data technology could be one of the enablers to collect these data into a distributed framework with physical storage on the commodity hardware. I am wondering about the data processing to extract the values for telcos. Today’s OSS/BSS are centralized into the software package with data and applications. Should we use in this future layer the massively paralelle processing closed to the data instead of the traditional OSS/BSS ?
Thanks Jian for the compelling comment. Your question almost feels like a trap, though I know it is not intended that way. There’s always the gap between the ideal architecture and the most practical architecture, either in terms of complexity or cost. If I stick closer to the ideal, but look at maybe a 5 year evolution, it seems to me that as has been the case with many other emerging business functions, like revenue assurance, it might begin as a bunch of little islands but ultimately it will develop into a continent of its own. In other words, I think I see more ‘small data’ than ‘big data’ right now. We seem to have many highly specialized analytics engines/tool, etc. growing out of OSS/BSS and other applications/processes/functions. In many cases, I suspect this is either because you have specific groups looking for specific results within CSPs but also because vendors understand they can sell more and increase their value with more analytics capabilities built off of their core products. If we’re really going to do ‘Big Data’ then I think we want to be able to look at all of the different types of data we might have available in a CSP, extract what we need, and process it in a place that is really geared to be a Big Data Operations center (for lack of a better term). I’m a space buff, so I’d think of this as mission control for useful business intelligence. I began to think about this model several years back when working through an analytics roadmap for Validas. It occurred to me that as valuable as our data about mobile usage was, it’s real strength could be in making other data even more valuable. In turn, that meant we’d want to plug as easily as we could into a client’s Big Data infrastructure. As it turns out, this approach resonated with folks in the data business – quants in the hedge fund industry and the big, metric-driven market research companies. So, I think this is very long winded way of saying I don’t think it ultimately makes sense to keep the processing tied up in traditional OSS/BSS. The real business insights that come out of Big Data need to be pursued for their own purposes. That doesn’t negate or eliminate the need for domain specific analytics that relate directly to certain OSS/BSS processes, it just serves another – perhaps more strategic- purpose. In the case of Acme Packet and Tekelec, there’s an awful lot of powerful data under the covers there that ultimately could drive the intelligence behind genuine personalization and real-time interactivity. Whether that’s for advertisers, content-aware second screen applications, and who knows what else coming down the pipe, it seems that’s where the whole digital services/digital content world needs to go (among other things). I think that’s revenue-facing/generating enough to warrant its own treatment and investment in proper and dedicated (in a conceptual sense; not the technical sense meaning ‘static’) resources.