Mixed projections for AdvancedTCA as new networks take hold
Analysts from IHS Global Research and VDC Research weigh in on the effect that technologies like Software-Defined Networking (SDN) and Network Functions Virtualization (NFV) will have on the market position of AdvancedTCA (ATCA).
After more than a decade of deployment in core infrastructure applications, the future of AdvancedTCA (ATCA) will be tested by fundamental shifts in network architectures. Technologies such as Software-Defined Networking (SDN) and Network Functions Virtualization (NFV) have begun to transform the way Communications Service Providers (CSPs) view the hardware elements that comprise the backbone of their networks. As David Laing, Senior Analyst/Program Manager at VDC Research in Natick, MA (www.vdcresearch.com) projects, this will result in mixed numbers for ATCA over the next five years as it competes with regular server platforms for the lowest cost per bit.
“There’s a war that’s going to go on between [ATCA] and the regular server platforms, particularly as things move to the cloud with SDNs,” says Laing. “ATCA’s advantage, in some cases, was the fact that you had all this redundancy right in the system, and when you really take a look at it with servers, once they move to the cloud the redundancy is in the cloud. That battle is still being fought, and in fact it’s just starting.”
“ATCA was designed as a distribution architecture, not a compute architecture, and it did the distribution quite well,” says Eric Gulliksen, Senior Analyst, VDC Research. And then you had folks like IBM with their BladeCenter coming along and saying ‘we’re the ones who are inside the datacenter. ATCA, that can be the interface between our stuff and the big wide world out there, and we can coexist.’ Then of course they tried to take their BladeCenter and turn it into a distribution architecture with mixed success; the BladeCenter T did not do nearly as well, according to our figures, as the original BladeCenter because it tried to get out into competing with ATCA. Now you’ve got the same thing going on in reverse; ATCA is going in and saying, ‘let’s go in and try to compete inside the datacenter.’”
Gulliksen adds that product upgrades from major Telecom Equipment Manufacturers (TEMs) have helped them regain market share, resulting in a general slow down for ATCA. His colleague, Chris Rommel, Vice President, VDC Research, also notes upturns from the likes of Dell and Hewlett-Packard telecom business units, noting that much of ATCA’s competition in the years to come will spring from general-purpose compute platforms.
“There is going to be the opportunity for ATCA and some of your extensive traditional networking systems going forward, and then there is going to be a growing percentage of the market also served by the more general-purpose compute platforms,” VDC’s Rommel says. “As opposed to paying for a $10,000 or $100,000 system, you’re buying just an array of general-purpose compute platforms that are good enough.
“There are kind of two sides of that coin that develop, where OEMs almost by organizational preference are going to try to determine what is ultimately going to provide them the cheapest cost per packet,” Rommel continues. “It’s almost a preference based on internal expertise and existing assets, but meanwhile you have an increasing number of large customers like Amazon and Google that are building farms of server appliances based on standardized hardware. So it will be very interesting to see how that dynamic plays out, but that said, that also comes in an environment where there is just the continued increase in the computational burden placed on networks that I don’t think there will necessarily be a loser per se, there are just going to be different systems obtaining a faster rate of growth than others.”
VDC’s Laing notes that one potential bright spot for ATCA is in Network Intelligence (NI) applications like Deep Packet Inspection (DPI) because it affords both the processing to inspect packets, as well as the switching capabilities to route them based on the information found. In those scenarios, ATCA “would seem to make perfect sense,” he adds.
Toby Colquhoun, Senior Analyst, Electronics and Media, IHS Global Research in Wellingborough, UK (www.ihs.com) predicts: “at a system level, ATCA will grow at a compound annual average rate of 10-15 percent from 2012 to 2017, although the growth rate in the near term is projected to be far smaller than this.” In similar findings, data from VDC Research shows a slight decline for ATCA CPU blades beginning in 2014 (Figure 1).
Customized COTS creating opportunities
While it is rooted in standards, recent trends show an uptick in customization of ATCA platforms. Laing explains that although the PCI Industrial Computer Manufacturers Group (PICMG) is making efforts to remove some of the benefits of massive customization, suppliers say that companies like Huawei are ‘bastardizing’ ATCA by attempting to extend it beyond what it was meant for. However, customization has reaped benefits for ATCA vendors that are willing to modify the Commercial Off-The-Shelf (COTS) technology, albeit rooted under the banner of an open standard, Rommel explains. Combined with integration strategies and target markets, this has contributed to the changing makeup of the ATCA ecosystem, Rommel adds.
“The dirty little secret in the marketplace is the level of customization that happens, because none of the vendors really want to admit it,” says Rommel. “Looking at the state of a couple of the largest vendors in the embedded space, there are some, like Advantech, that are gaining market share because they are more willing to do customized derivatives of their various systems and have fewer SKUs, whereas other large vendors based in Europe aren’t as willing to do as much customization and historically have had too fragmented, too large a portfolio of products.
“One of the interesting dynamics at play looking at the ATCA market just from a competitive standpoint is the successes and failings of some of the players in the marketplace,” Rommel continues. “Largely that has to do with some of the technology acquisition habits of some of the largest Tier 1 TEMs, and how some of those different companies have, to varying degrees of success, tried to pursue that opportunity. You look at Radisys and Emerson in particular – both have done pretty poorly over the past handful of years. One of the biggest issues as a whole for those two was initially both had pursued the Tier 1 TEMs too aggressively, and now what’s happening is the TEMs are passing on some of the cost pressures they’re facing, and they’re only looking for boards as opposed to more complete systems. That has caused the margins available for the likes of Emerson and Radisys pursuing those marketplaces to really erode.
“There are other vendors that take a different approach, like Advantech or ADLINK as an example, that have focused on Tier 2s in the networking space, that do less integration in house so they can do higher value systems, higher margin systems, and likewise have been putting more effort and had more success expanding the reach of ATCA into some additional vertical markets,” VDC’s Rommel says. “Advantech has even spun out its own derivative standard product, eATCA. That is one of the really interesting things that has happened in that marketplace, just how some of the original stalwarts have been stumbling. Clearly Emerson got to the point where they divested a large percentage of their stake in the marketplace.”
Other PICMG standard projections
PICMG specifications beyond ATCA are seeing growth due in part to their applicability beyond the core network. IHS Global’s Colquhoun sees MicroTCA (µTCA) as one such platform to expect more growth from than its larger counterpart, be it from a smaller base.
“The market for µTCA systems will grow faster than the market for ATCA components,” Colquhoun says. “The reason is that more components are used outside of traditional communication applications. µTCA will be increasingly used in higher growth sectors like medical, industrial, and High-Energy Physics (HEP).”
In addition, both VDC Research and IHS Global Research show the original CompactPCI (cPCI) specification remaining viable, with IHS predicting a significant revenue shift towards 3U platforms by 2017 (Table 1). Outside of communications, the Aerospace and Defense market is helping sustain original cPCI technology. Metrics for cPCI Serial are still too limited to glean meaningful insight.
IHS Global Research