Oracle World 2013 finished up a couple of weeks ago, and I’ve had a bit of time to reflect on the event. On the Wednesday of the event I spoke at a luncheon hosted by the former Acme Packet Group, which was acquired by Oracle earlier this year. I hadn’t been to Oracle World in a number of years and I wasn’t sure what to expect given the fact that historically Oracle and communications went together about as well as Larry Ellison and Bill Gates.
However, things are changing and everyone is jockeying to move into other markets. That’s why Cisco sells servers and HP sells networking gear. Since Ellison chose to go watch his ship race rather than show up to his own keynote, I have no idea whether he was going to mention communications or not, but make no mistake, Oracle has moved into communications and is here to stay.
Much of the focus of the Acme Packet lunch was on SIP trunking, which was highlighted by a large customer of theirs that had recently migrated the company to all SIP trunks and talked about some of the best practices regarding the migration.
One of the points brought up was that if organizations are going to migrate to SIP trunking, then it’s an absolute must to re-architect the environment and aggressively consolidate down the number of trunks. With legacy voice, every location that has a PBX or a key system has its own dedicated trunks. Even small locations may have a couple, for redundancy purposes, meaning the utilization of the trunks is typically very low. Any company doing SIP trunking should get rid of local trunks and consolidate down to a few centralized trunks out of the data center, and then use the company WAN to connect the local branches to the data center.
A key to this is to make sure the network is robust and redundant enough to handle the voice traffic. MPLS can be used to create a class of service to ensure voice quality, but this particular customer was actually running dual networks to ensure 100% uptime. One of the common thoughts regarding UC is that organizations will save money by going from two networks (one voice and one data) down to one data network. However, from a best practice standpoint, it’s actually better to run two data networks to ensure redundancy since the call control should be centralized. Ideally, companies would have one traditional MPLS network as the primary and then use some kind of wireless service, such as 4G for a backup. If the network is architected correctly, companies will realize higher levels of uptime and quality than with a traditional voice. Some may look at this and think that running a totally redundant network defeats the purpose of moving to UC in the first place and obviates some of the savings. To me, that’s short-sighted. There’s so much money to be saved with the SIP trunking migration that even if the networks wind up costing more, it’s still well worth it.
Looking past just SIP trunking, there was a fair bit of discussion of “communications enabling” applications, and here’s where Oracle and other IT-type companies can have a big impact. I sincerely doubt that Oracle has moved into the communications space to be yet another UC vendor. The opportunity for Oracle or other IT-focused companies isn’t to build UC platforms but rather to use UC as a way of building innovative applications that can connect customers to information faster. I know that’s the talk track from the traditional UC suppliers, but the focus is different. For Oracle, communications enablement is the primary focus, rather than something that comes after the best widget is built.
Looking out to future Oracle Open World events, I do expect to see communications become a bigger theme at the event. Just don’t look for it to be overtly out there. Rather expect to see topics such as SIP and WebRTC get embedded into other topics that bring the industry closer to finally having some real communications-enabled applications.