[Previous entry: "Mojo mosh networking"] [Next entry: "See you @ Starf**ks"]

08/07/2004: "Vivato performance for under $600, a tale"

And now the story must be told.

In March of 2003, it was becoming more and more evident that Vivato's first product was not going to deliver on the promise. The situation was so bad that I wrote an absolute flamer to the then-CEO, Ken Biba, explaining some of the real-world issues with the 1st generation product, and proposing a new and different product. That memo resulted in what eventually became Vivato's second 'switch' product.

Still, the internal fans of the first product perservered. Whenever anyon would raise technical concerns, or a concern about the manufacturability of the first-generation product, the answer was always refrain; "We can do it".

Bob Conley tried to get me fired for leading Vivato down the commodity path. He didn't like the second generation product. Reports are that it is meeting with some success, despite Bob's concern. I hear that sales of the second generation switch (which is still indoor-only) sales are picking up, and that a backlog exists. This is better than the situation with the first-generation switch, where the 'backlog' is the warehouse. Several hundred still exist, like forlorn puppies in a window of a pet shop. Built and paid-for, still stitting in boxes, waiting for a customer. VC dollars hard at work, transformed into inventory that nobody will buy.

There were still real problems with the product operating in its "three simultaneous beams" mode. No technical arguement seemed to sway the deciples of the third beam from their "We can do it" mantra. Both I and others made repeated whiteboard arguments why "three simultaneous beams" wasn't ever going to work.

Three channel (three simultaneous beams) had become a religion, and the zealots had the floor.

In April of 2003, I proposed a test. Those who knew me, and my agenda thought I was being as wicked as they'd ever seen. Phil, my boss at the time, actually labeled me "wicked". Those who were members of the opposing religion, the Agilent/metropolitan broadband faithfull were about to get a wake-up call. What I proposed was this:

We take a fresh-built, tuned-as-they-like, 1st-generatin Vivato switch, and pit it against something that I would come up with "off-the-shelf". I had a budget of $2,000.00 .vs the $9,000.00 to $14,0000 Vivato product. The Vivato product had to either have twice the range, or twice the throughput at the same range.

Everyone agreed that the Vivato switch would need at least this much advantage to demonstrate its value proposition.

So, I borrowed a Cisco AP350 (we had several internally), and went googling for a set of high-gain waveguide antennas. I remember simply hoping that someone had made a commercial variant of Trevor Marshall's waveguide for the 2.4GHZ ISM band. On page 3 or 4 of the 2nd or 3rd Google search I tried, I hit pay-dirt. I found a company named Securawave who had just started shipping a waveguide antenna in two different gains, at the time, 15dBi and 18dBi. (Securawave have since upped the advertised gain on these antennas to 18dBi and 21dBi.)

Not wanting to expose Vivato to a competitor, I checked with my manager, and once approval was given (Ken Biba also chimed in), I purchased them through Netgate. I purchased 4 of the 8-slot, and two of the 16 slot "multi-beam" antennas from Securawave. We really did have early units, when they arrived none of the antennas had a serial number higher than '10'. These antennas cost $250 to $325 each, the same then as now.

At the time, a Cisco AP350 could be had for approximately $725. I also purchased a set of rp-tnc to N cable assemblies from Hyperlink, these may have been $20 to $25 each. At $1300 total, I beat the cost target.

Several people inside the company though I should re-use some of the ridgeguide antennas off Vivato's early prototypes. I considered it, but I decided that this violated the "off-the-shelf" aspect of the experiment. I didn't want anyone coming back to ponder that getting a waveguide antenna built would stop someone.

When the testing was over, the results were discouraging. The Cisco AP with the waveguides had about 20% more range than the Vivato switch. The Vivato could only manage about 25% more aggrigate throughput, and thus, the Vivato product failed the test. Vivato was beat by less than $2000 of off-the-shelf equipment.

But we shipped it anyway. Vivato's future was on the line, there was a round of VC financing to close. There was some softening of the "three beam" position though. The Vivato switch now ships as a "two beam" product. The problems are still there, but they aren't as noticible in a 'two beam" configuration. Zealotry won, as it often will.

Those of you who want to clone the system that beat Vivato can easily do so. Just over a year later, one can now come within 3dB of the (claimed) first-genration Vivato performance for less than $600. Given the heavy self-interference inside Vivato's first-generation switch when operated in anything but the "single-beam" mode, the performance may well be beter. As always, your milage may vary.

The Securawave antennas are available (and we still sell them at Netgate), and the guys at Securawave report that they best Vivato at nearly every opportunity.

Additionally, and at my urging, Poynting has developed and started manufacturing their own waveguide antenna. Netgate will be carrying this antenna as well. I have an early sample here in Hawaii, and its a sweet little unit. Couple either up to the HS3000, and you'll have a less than $600 solution that will run with Vivato's $14,000 outdoor solution or their $9,000 indoor switch.

In previous posts, I've pointed out the flaws in attempting metropolitan deployment in the unlicensed band. Apparently my words have received some amount attention both in Spokane and elswhere. A reporter for the Spokane newspaper, the Spokesman Review, called to ask me several questions. He's queing up a story, and I apparently play a part.

He's asked other locals for their opinion, with 180 Network's Chad Skidmore responding that while I have several accurate points, he thinks that Vivato can fix the problem "in software". Unfortunately, Mr. Skidmore is mistaken. The problems happen before any layer of software can run. They are indemic to the design of the receiver system.

The reporter will also likely paint me as a disgruntled former employee. While it is true that I am disgruntled, and a former employee, my reasons for continuing to post these articles are my own. And on that subject, two things:

First, I am far less disgruntled now that I have cleared the county line. Spokane sucks. Nuff said.

Second, my posts about Vivato aren't so much "get back" but rather "fair warning" to any who read to be careful. Investigate, ask questions. its your money, after all. And should Don come waltzing into your company next, perhaps you'll know more than we did.

Don, after all, is the man who said he fired the former VP of Sales because he "didn't know how to sandbag". In sales, "sandbagging" is the unethical practice of artifically holding the quota system low, or just grazing over quota in a given time period so as not to fall short the next, or otherwise advantage oneself of the bonus-based quota system. Sandbagging is not a behavior you want in your VP of Sales. Yet Don announced this as the reason for sacking Greg. Greg is now enjoying great success at BelAir and oft claims that its refreshing to have a product to sell that works as promised.

But back to money, Vivato held another layoff yesterday (Friday), with it went Vivato's last hope of any new product development. Vivato laid off Siavash Alamouti and eight others, as well as 8 of the 12 contractors from Mindpath. Siavash and four of the eight were amoung the brightest technical minds that ever worked at Vivato. One of the people affected sent this message back to those who remain:

I still remember the first standards meeting after we announced
Vivato, when everybody used to crowd around me, every singe break
period, like I was a rockstar, and I kept repeating, "No changes to
client card!" J Boy! - What fun! it has been an exciting two years for
me. I changed my definition of feasibility couple of times over and
over, leant humility, tasted ignorance, now its time to ride on.

The end result of this is that there won't be a lot of new product development coming out of Vivato. The pile of cash is too small, and now the only people who could have designed a product that worked have been dismissed.

I am told that part of the reason that Siavash is gone is that he was at odd's with Don Stalter's plans to turn the company into a metropolitan deployment company. Siavash is, in a word, too honest for Don to continue to endorse his presence. Siavash knows that Vivato's current products won't work in metro-broadband deployments. He must had said so, and Siavash got the axe today. In Stalter's company, there is only room for Stalter's opinion. The man sneered at me once when I countered his opinion with fact, and I do mean the lips thinned, full teeth barred "snear" move. It was animalistic. I would have made a fist and smacked hard across that thin little mouth of his, but for the simple reason that I abhored giving him an excuse for all that has happened since.

I am sure that the world has not heard the last of Siavash. As 802.11n continues to gain traction, the man who invented "Alamouti's technique", the single most efficient MIMO system possible, is sure to find a new position. If you're a VC, and want a chance at owning the MIMO/802.11n market, let me know, and I'll get you in-touch, but move fast, Siavash already has interviews lined-up.

The rumor mill says that in order to keep the seat warm, Don Stalter is hiring Ron Rudokas, formerly of Vistard. Those of you reading this from behind the wall, or who are fellow escapees will recognize the name. Vistard was founded to exploit Vivato's advantages in deployment of metropolitan broadband. When the Vivato product didn't work, Vistard failed to obtain funding, and the whole thing went nowhere. The way it looks from here, Mr. Rudokas is joining Vivato to pursue the dream, but Mr. Rudokas is no replacement for Mr. Alamouti.

To fill other positions, Don Stalter continues to hire his friends from former companies. Other new executives to Vivato are Steffen Rasmussen, as Sr. VP of Engineering, and Doug Smidl, formerly of ADC Telecommunications, Don's employer prior to Terawave. Given this behavior, I now openly question if Vivato's VCs are watching their investment, or if they have given up.

More yes-men dressed as chiefs, but fewer indians by the day. Not a formula for success I've read about in any of my favorite business books.

Don apparently promised the company Friday that Vivato would IPO, and the remaining employees would all be rich. I wish this could be true, but the market would need to value Vivato at more than $100M before any of the common shareholders would receive a dime. Given what I know, the company lacks the revenue to justify that kind of valuation and I find it suprising that Don would make this promise, given that it is quite probably, "actionable".

From what I hear, without additional cash, they'll be lucky to be in business in 12 months. If you're still there, the time to get out is now. Its over.

I've now said all I ever intend to say about Vivato, and Don Stalter. You won't hear more from me on the subject unless they foolishly decide to attempt to sue me over my postings, or a small annoucement here when they finally fail. I have but one condition, don't use ex employees to spin your culture. Three of the four people pictured no longer work at Vivato. Clean it up, Don. (And he did.)