* Branches to turn to QoS to aid VoIP rollouts
endif; ?>Our recent discussion of frame relay over DSL gave a relatively strong endorsement of the technology. In fact, even our discussion of mean-time-to-repair issues indicated that this shouldn’t be a major problem in most cases.
Market research firm Ashton, Metzler & Associates recently surveyed the subscriber base at Webtorials.com to learn about trends in branch-office networking. The firm concluded that, in general, within the next year, branch- office networks will be required to support a number of new applications that will demand a very robust network infrastructure.
According to the “Branch Office Networking State-of-the-Market” brief, authored by consultant Jim Metzler, the biggest application change to be seen in early 2005 will be in the number of deployments of VoIP in branch offices. VoIP ranked third in application deployments, following e-mail (99%) and backup/storage (87%). A full 79% of respondents said they will be running VoIP in branch offices in early 2005, representing a 295% increase over late 2003.
It comes as no surprise, given the VoIP trend, that the most common networking capability that the researcher anticipates will get deployed at the branch during the next year is quality of service (QoS). About 34% of the survey respondents intend to deploy QoS based on queuing or on MPLS in the next year.
Metzler expressed surprise, however, at the fact that there was less interest in compression; only 10% said they intend to deploy this bandwidth-optimizing capability.
One reason might be that, often, technical issues emerge if you try to combine heavy-duty compression technologies (those that yield up to 10:1 throughput improvements) with QoS using different vendors. Usually, if one vendor’s device compresses traffic and another’s attempts to shape it, the functions won’t work together. This is because shaping devices can’t identify data that has been compressed by another vendor’s proprietary algorithms. (Similar issues arise with encryption and QoS.)
To fix the QoS-compression conundrum, you might have noticed that vendors have now integrated these functions in their devices so that they recognize one another. Packeteer, known for its monitoring and shaping strengths, for example, added compression to its PacketShapers a little over a year ago. Similarly, Expand and Peribit, which got their start with sophisticated compression devices, added shaping capabilities to their products last year.
The good news with these vendor moves is that you can do both functions. The bad news is that if you prefer one vendor for its compression strengths and another vendor for its traffic-shaping strengths, you’re more or less in a pickle.




