The decision to buy a three-stream access point (or likely, a whole lot of these) in an enterprise environment is more complex than throughput alone. You need to begin with overall IT requirements, objectives, planning cycles, and operational strategies.
So, starting at the beginning, the greatest value of advanced Wi-Fi technologies is in their ability to handle the increasing number of both users and wireless clients, as well as the increasingly number of applications on the corporate LAN, including those requiring time-bounded services.
In other words, an improvement in multi-user capacity inherent in three-stream implementations, as opposed to single-client throughput, should be the primary motivator for considering this alternative, realizing, of course, that the vast majority of clients will likely remain single- or two-stream over the useful life of any access point purchased today.
Higher throughput, then, is of primary value in maximizing overall channel utilization as measured by the number of users or sessions simultaneously satisfied - simply put, the faster and more reliably a given station sends and receives its desired bits, the more time and thus more capacity remains for everyone else. Even time-bounded traffic can benefit, as fewer mangled or lost packets result in higher voice and video quality.
It may very well turn out in any given case, based on a careful review of network management data, especially with respect to utilization trends, that a two-stream (nominally, 300Mbps) access point will be more than adequate, given estimates as to utilization towards the end of a given planning horizon.
If this is on the order of two years or more in the future, it is again very likely that 802.11ac will be the preferred upgrade path from two-stream 802.11n, rather than an interim upgrade to three-stream .11n. Given that we expect little price premium for 802.11ac access points over two-stream .11n on the order of two years from now, such a strategy may make the greatest economic sense.
If, however, current capacity utilization or related trends indicate the need for more capacity sooner than that two-year horizon, the next step will be to check with one's system vendor regarding product availability and any issues (which should be minimal) in running a mixed-mode two-stream/three-stream infrastructure.
Any decision to deploy a greenfield three-stream-only infrastructure will likely be based on price alone - we expect most enterprise-class systems vendors to continue to price aggressively, so the price premium of three-stream .11n over two-stream will likely be inconsequential. And we expect essentially all enterprise-class WLAN system vendors to have three-stream products in their product lines over the next six months.
We do not, by the way, expect four-stream (600Mbps) access points to have any real impact on the enterprise market - this direction will clearly be eclipsed by 802.11ac. But some four-stream products will still appear, particularly in carrier and service-provider applications.
Buyers of residential/SMB-class products, however, should consider only three-stream products right now. Again, there is no price differential of note, and the additional capacity here will benefit especially gaming and video applications. And, as we saw in these tests, owners can legitimately claim to have the best available technology at their disposal.
But a final point: throughput alone is not suitable as a basis for a decision on a particular product. There are dozens of other features and corresponding benefits that must be considered, from overall system architecture to non-disruptive growth to support for security, integrity, management, integration, and much more. So, just as one wouldn't buy a switch or router based only on performance specifications or the results of throughput-centric benchmarks, such advice continues to apply to wireless LANs as well.
Read more about anti-malware in Network World's Anti-malware section.
This story, "Tips for Buying a Three-Stream Access Point" was originally published by NetworkWorld.