Valley Wonk: Juniper's Shopping List
One Wall Street visitor at Juniper's recent analyst day had a good solution. "You know what you should do?" the analyst told Juniper execs. "Just buy something, just to shut everybody up."
It's not a bad plan. Clandestinely set up two guys in a garage, take their picture next to a blinky-lights box that says "Layer 4-7 Ethernet Wireless Security" on the side, and announce the acquisition of AirEther4-7 Inc. – for an undisclosed sum, natch. Poof! Every merger rumor comes true at once.
CEO Scott Kriens went for the humor tactic instead, opening the analyst meeting with a joke about a hostile takeover of Oracle (see The Price Is Right for PeopleSoft if you don't get it). But that was it, and it left analysts to keep obsessing about the acquisition thing.
One source claims Juniper has three acquisitions on its wish list: something in Ethernet, a technology to "fill a hole" (presumably wireless), and a "mid-market" company. Our guess is that the last one is a place-holder for the budget.
Luckily, the most plausible merger rumors fall into three categories. So, let's run with the idea: Where should Juniper's merger scouts be looking, if they aren't already?
Yes, Kriens hates the idea (see Juniper Spikes M&A Rumors). But it isn't going away, because many believe Juniper can't succeed in the enterprise without matching Cisco Systems Inc.'s (Nasdaq: CSCO) breadth of product. That means Juniper needs an Ethernet switch to sell alongside its routers.
The problem is the lack of easy – or cheap – candidates. Force10 Networks Inc. is the only private company that seems likely, and it's going to be too expensive, mostly because the VCs, who have pumped hundreds of millions into it, want to see either top-dollar or an IPO(see Force10 Networks). That leaves Extreme Networks Inc. (Nasdaq: EXTR) and Foundry Networks Inc. (Nasdaq: FDRY), both public and expensive – market caps of $770 million and $1.4 billion, respectively. Sources have said Juniper pinged Foundry on the idea last year but got rebuffed by Foundry CEO Bobby Johnson.
Juniper probably does want a hand in this market but doesn't like any aspect of it. To defray the high acquisition costs, maybe Juniper should just build its own stuff. A chip acquisition would be cheaper than an Extreme buy. Or, Juniper could just use off-the-shelf chips as the basis for Ethernet switching cards.
But this still doesn't solve the margins problem. Given the lack of easy options, Juniper might be best off stepping into a reseller role, offering Ethernet switches through a partner.
Wasn't this a hot topic a month ago? Once Cisco grabbed Airespace Inc., everyone expected Juniper to dive for a wireless LAN switch company. And in fact, Juniper reportedly put in a bid for Aruba Wireless Networks and talked to the less expensive Trapeze Networks Inc. (see Sources: Juniper Eyeing Trapeze).
Merger talk seems to have died down here. The companies here are cheaper than the Ethernet choices, but some might be seeking better options while WLAN is still a hot topic. Aruba, in particular, might have its eye on going public.
And again, margins might be a factor. Wireless LAN switching is a separate appliance for now, but it's going to become just another feature inside a switch or router. That means the WLAN switch's future is one of cost-cutting and integration, with chips and software already emerging to help things along (see Broadcom Integrates WLAN Switch and NextHop Reveals WLAN Plan).
Wireless might end up being another reseller category for Juniper. That would give the company a WLAN story for the moment while developing its own WLAN switch-on-a-card for later.
Layers 4-7 processing
Higher-layer processing presents better margins than Ethernet or WLAN, and plenty of startups are working on this technology, presenting lots of options. It's no wonder this has become the latest pet theory.
The logic goes deeper, too. This technology would make a nice extension to NetScreen and to the Infranet Initiative, a Juniper-supported effort to drive standards for a more intelligent public network (see Juniper Buys NetScreen and Juniper's Infranet Takes Baby Steps). And Kriens keeps talking big about traffic processing, saying it's central to Juniper's role in the network.
Juniper officials point out that they already have this stuff. But Cisco is amassing some new tricks with technology acquired from the likes of P-Cube (see Cisco Plucks P-Cube for $200M and Cisco Rules L4-L7 Switch Market). Juniper ought to counter.
F5 Networks Inc. (Nasdaq: FFIV) gets held up as the example here, but plenty of others could be had for cheaper, including NetScaler Inc., Peribit Networks Inc., and Radware Ltd. (Nasdaq: RDWR). Not to bias the judges, but NetScaler happens to be one of Light Reading's top picks to get acquired or go public – and it occupies that coveted No. 10 spot on our Top Ten Private Companies list.
Given the options, it doesn't look as if Juniper will go through with three acquisitions -- not from this pool, anyway. A Layers 4-7 buy looks like a slam dunk, but the company has to make some tough decisions in Ethernet and wireless.
In the end, it seems clear that Juniper is set on trying to create a new kind of enterprise company, one that delivers a higher-performance network.
The other option? If Juniper holds off long enough, it could just be another enterprise vendor with some gaping holes in its portfolio.
— Craig Matsumoto, Senior Editor, Light Reading