By David Newman, Network World Lab Alliance , Network World , 04/28/2008
1. What applications need to be unified communications-enabled?
Many enterprise IT organizations divide responsibility for networking and applications into two groups. Unified communications requires both application and network infrastructure support, so that means getting both groups on board before deployment. Enabling unified communications support may involve a little or a lot of work, depending on the application. For some messaging and VoIP applications, presence support may already be built-in.
Other applications – such as databases, CRM packages and especially custom-developed transaction-processing programs – may lack presence support altogether. Identifying which applications will need unified communications support, and getting the appropriate development help if needed, is a necessary first step.
2. What existing network infrastructure supports unified communications?
Passing along presence information in a unified communications-enabled network usually means supporting new protocols and/or adding infrastructure services. With location-based unified communications services, for example, DNS servers may need to be updated with SRV (service type) and LOC (location) records. Other infrastructure services, such as those for e-mail and instant messaging, may need to be extended. Presence information (for example, to move availability status between voice and instant-messaging systems) can be added via SIMPLE or XMPP.
Security is always a concern whenever adding new pathways through the network. Most unified communications systemscan use centralized directory servers such as those based on Lightweight Directory Access Protocol or Microsoft Active Directory to authenticate users or processes seeking to communicate. And firewalls and other security devices will need to be reconfigured to support new messaging protocols – not just SIMPLE and XMPP but also any proprietary messaging protocols such as those used by AOL Instant Messenger, Yahoo Messenger, and the like.
3. What about presence for existing PBXs?
Telephony systems have supported presence services for decades (think do-not-disturb and call-forwarding features), but that doesn’t necessarily translate into unified communications support for messaging, video and other data types. It’s relatively easy to add presence support to open-source platforms such as Asterisk or OpenSER. For proprietary PBXs (even those based on derivatives of the ITU’s H.323 specs), more work may be needed. Check with your PBX vendor about adding support for presence protocols such as SIMPLE and/or XMPP.
4. How will you manage unified communications?
Given the huge interest in collapsing services into centralized data centers, it may make sense to add and manage unified communications from a central location as well. During trials, for example, it’s far easier to locate a single Jabber server at a central data center rather than to distribute servers throughout the enterprise and then try to keep them all in sync. Centralized management also can simplify change management. Alternatively, unified communications can be totally outsourced as well; professional services groups at major PBX vendors such as Avaya and Siemens offer unified communications as a service.
5. What about unified communications support for future applications?
Because the concept of unified communications is so broad, it’s quite likely additional applications and network services will eventually be needed, well after initial deployment. Experience during initial trials may be instructive: If it’s difficult to integrate, say, LDAP support into one application, then adding it to 10 applications could prove far more challenging. A post-mortem after initial deployment will help you understand which of unified communications’ many parts do and don’t work in your organization.