Thursday, March 3, 2011
HP E-Series Mobility Portfolio
HP has launched a new series of access points through a combined development effort with the HP/Colubris development teams. The new access point model numbers are the E-MSM460, E-MSM466 and the E-MSM430.
HP's goal is to bring a 'single pane of glass' management capability to the wireless and wired networks through integrating the HP Mobility Manager 3.10 into the existing IMC solution. Mobility Manager can be a plugin to an existing PCM+ installation.
The biggest news to me was the AP MSM466, which is capable of concurrent radio operation in the 5GHz band. This allows the access point to increase the channel capacity to double the supported client count in high density deployments. This published statistics for this access point indicates a maximum performance of 450Mbps per radio. Using two 5GHz radios in an access point is interesting, but there are still a lot of 2.4GHz clients in use on most every WLAN. Having all your clients in a specific area being only 802.lla devices may be a reality for some enterprise deployments, but I'd bet that most have a wireless client mix that can't be controlled or influenced by the IT department.
The HP mobility line can support different modes of operation - AP, Mesh and Monitor (packet capture) modes. The new features of the HP mobility hardware producte line are standards based beamforming (explicit) and band steering. There was no mention of the ability to do spectrum analysis with any of the HP access point offerings. The lack of spectrum analysis as part of their product offering does not allow the HP mobility portfolio to identify sources of interference. The HP mobility product line can only adjust the power and channel of the access point in reaction to sources of interference.
I thought the slide showing the comparison of the HPMSM410 and HPEMSM460 to the Cisco AIR-LAP1142N-A-K9 access point was a little misleading.
Mostly since the TxR:S numbers for each of the access points are not clearly stated on this slide. The Cisco 1142N access point is a 2x3:2, and the HP MSM410 is a 3x3:2 access point.
I found it interesting that the MSM410 performed only slightly better than the 1142N even though the radio in the MSM410 has three transmit and three receive antennas. The comparison difference is marked between the E-MSM460 and the Cisco 1142N due to the fact that the E-MSM460 is a 3x3:3 access point. The metrics on this chart show the E-MSM460 providing 150Mbs of throughput at a distance of 230 feet from the access point. This works out to be one access point every 1400 feet. If this distance is to be used as the gauge for the cell edge, that's a pretty dense access point deployment!
One thing I found of note was the ability of the access point to be changed into an autonomous access point just by changing the operating mode on the access point from the controller. You're not required to change the code running on the access point in order to make the access point function independent of the controller.
The HP mobility solution does not use the CAPWAP standards-based protocol for their controller based solution. HP uses a proprietary wireless protocol that is based on IAPP and using OpenVPN with UDP tunnels in order to simplify network connectivity on LANs using NAT.
This mobility announcement from HP will be great news for existing HP mobility customers, but I am doubtful that customers with an already deployed WLAN infrastructure will find enough compelling features to make the switch to the new HP E-MSM product line. However, some customers may require the cost benefit of the next day replacement that is part of the HP lifetime warranty.
Jennifer, thanks for your time with HP Networking. I'd like to clarify a few things that you have pointed out in your article above.
ReplyDelete1. I'd like to say that even though dual concurrent 5GHz may not be suitable for all indoor deployments as yet, that would definitely be the model moving forward a few years from now. Secondly, dual concurrent 5GHz is immensely valuable for outdoor building-to-building links, today.
2. Spectrum Analysis capability is built into the existing hardware, but has not enabled as yet.
3. Your point about the number of transmit/receive antennas and spatial streams is well-taken. We are primarily trying to show the benefits of the new technology here.
4. As for your comment on range, I am not quite sure how you arrived at those numbers. However, one comment I would like to make is that the range on these new access points is no less than the existing 2 spatial stream access points out there. It is only better, and in some cases up to 50% better.
5. Not only are these new high-performance APs valuable for new install, existing customers currently struggling with bandwidth issues on their wireless networks will greatly benefit with the improved capacity and range of these high-performance access points.
I would like to know what shielding between radios HP has implemented since they are supporting dual 5GHz concurrent operation. With 2 radios less than 10 feet apart the strong signal strength of such a close neighboring radio on an adjacent channel will cause tons of interference unless properly shielded.
ReplyDeleteThe IEEE standards only dictate a relative signal attenuation (in dB) at frequency distances away from the main carrier frequency.
Thanks,
Andrew vonNagy
Hi Jennifer,
ReplyDeleteAs always, you amaze me. I'm looking forward to seeing you a WFD in San Jose.
I would like to know the answer to Andrew's question and one of my own...
For these units to run on .3af PoE, they have to have a minimal complement of hardware (1 gig port, 2 3x3:3 radios, minimal RAM, small CPU). If they do distributed forwarding at high data rates, they will have to apply security (firewall, encryption/decryption, etc) and QoS policy at the AP. With no encryption offload processor and a small CPU (required due to minimal PoE power), how will they handle high throughput in a distributed forwarding scenario?...especially with a high density of clients.
Also, due to .3af PoE, they had to have used standard-power Atheros (or other) radios, not high-power radios, so their range will be lower than those vendors who opted for high-power radios.
Thoughts?
Devin Akin
Chief Wi-Fi Architect
Aerohive Networks
http://blog.aerohive.com/blog
Anil,
ReplyDeleteI went back and double checked my math and found my square footage calculation was *way* off. I corrected the math and the chart indicates the data coverage was for an area of 166,106 square feet. That does seem like a very large area to be covered by a single access point even in a testing setup. I wonder what the testing environment looked like and if that square footage shown by the chart is accurate. By this I mean the edge of the testing area in the chart seems to be 230 feet.
Andrew, the dual concurrent behavior is supported on the MSM466 model that supports external antennas. The externally connected antennas can be physically separated to achieve the required antenna isolation.
ReplyDeleteThanks
Anil Gupta
Technical Product Manager (Mobility)
HP Networking
1. Devin, 802.11 encryption is handled in the MAC/basedband chipset, and not by the CPU.
ReplyDelete2. Upstream/Downstream QoS is handled by the MAC/baseband and the host processor, which it can handle very easily.
3. The transmit power is limited by the regulatory domain settings, because FCC and most other domains impose a MIMO penalty. It is not currently limited by the radio's transmit power. That limitation will apply to all vendors, not just HP.
Thanks
Anil Gupta
Technical Product Manager (Mobility)
HP Networking
Jennifer, I am not sure if I would fully agree that the "pie r square" method (where r=230 feet) that you have used in your new calculation to arrive at the 166,106 sq feet is the right method to calculate the square feet area of wireless coverage.
ReplyDeleteWe placed a laptop that was 230 feet away from the AP, that was mounted on the wall, to report those numbers.
Hope that clarifies.
Thanks
Anil Gupta
Technical Product Manager (Mobility)
HP Networking
Which Atheros chipset these AP's are based on?
ReplyDelete