Policy Revenues Climbing, but Hazards Loom
There aren't too many telecom equipment sectors that have been recording double-digit increases in revenue for five years or more, but the policy management sector is one of them.
Yet the news isn't all good: In our latest Policy Control & DPI Market Tracker, we note a significant slowdown in growth, from a whopping 30 percent in 2012 to around 17 percent in 2013. A key factor: The initial market for policy servers and deep packet inspection (DPI) gear is pretty much saturated, at least in the wireless sector.
So is the party coming to an end for the sector's many players? Not quite, our forecast predicts. We expect the market to nearly double from $1.45 billion in 2013 to $2.8 billion in 2018, a CAGR of 13 percent -- still pretty respectable by the standard of the overall telecom market.
Despite that initial saturation, there are many new use cases driving a second or third round of spending by telcos, building on initial deployments or adding new suppliers: Many report that they have replaced incumbent vendors in some engagements.
Yet despite this broadly positive view, we wouldn't be doing our duty to clients if we didn't point up the risk factors that could derail this market.
In the short term, those risks arise mainly from the ongoing transition to a more complex policy environment that supports a range of new cases. For the most part, these enable telcos to offer a wider range of service packages and options, such as packages built around specific applications, shared usage plans, temporary passes, and so on. Support for VoLTE, telco WiFi, end-user security, prioritization under congestion, M2M, and sponsored data are also emerging on wish lists.
The good news is we've clearly moved on from theory to practical implementation with some of these ideas, with both vendors and operators reporting a wider range of new packages actually being deployed, especially in emerging economies. Here, it's not unusual to see 20 or more plans, with Facebook-only, Google Zero, and daily data passes among the options.
The not-so-good news is that these plans require a different kind of business justification, based on new revenues rather than lower costs -- which is often harder to justify. Product marketing and development must be engaged, and that's not always a simple matter. Moreover, the new policy use cases often involve some level of integration with charging systems or other telco IT gear. And that's proved equally problematic, our operator surveys show. One result: Vendors report long lead and deployment times, slowing revenue recognition.
In the longer term, meanwhile, the risks are primarily associated with the transition to virtualized networks, including networks based on SDN. These introduce a whole range of uncertainties that are hard to read at this point. On the whole, we think that the role of policy management will be retained and perhaps enhanced in this environment: That was certainly the view of operators we polled last year, with four out of five rating policy as "critical" or "important" in an SDN network. And in recent conversation, ONF director Dan Pitt concurred, saying that policy software would likely act as a bridge between business rules and network control software.
Yet we should not underplay the tremendous changes that both NFV and SDN could usher in, and we shouldn't discount the possibility that new ways to control and differentiate service flows might emerge that supplant or even replace conventional approaches using policy servers and DPI. Nor is the revenue model in a virtualized environment settled: In a highly competitive market, could some vendors chase market share with virtualized or cloud-based products that disrupt conventional pricing?
On balance, we still think that policy software will play a key role in this emerging environment, and that revenue will continue to climb. But nothing is certain in our industry, and we'll be keeping an especially sharp eye open for the unanticipated developments that follow in the wake of every big new architectural shift.
ó Graham Finnie, Chief Analyst, Heavy Reading