Fool HQDiana Corp: Scam or Reality?

[EDITOR'S NOTE: Today, Rogue looks at Diana Corporation, whose stock rocketed to stratospheric levels earlier this year before falling back dramatically. Is Diana's technology of real value? Or is it just a shell?]

Peter Lynch, Warren Buffett, and most other successful investors insist that one of the best ways to lose money in the stock market is to invest in technology stocks you don't understand. In the midst of investor fads, this axiom may seem like merely quaint folk wisdom. For nearly a year, anything even vaguely connected to the Internet soared, and speculators savvy enough to ride the wave have made some fast and easy money.

But woe to those who believe the hype even if they can't understand the technology. When the wave crashes, your portfolio may well be nothing but a shark's next meal.

Investors in Diana Corporation (NYSE: DNA) know the feeling. The stock rode a wave of manic momentum buying from a 52-week low of around $10 a share last fall to an all-time high of $114 late this past spring. [See Diana's performance for the year: DNA Quotes: Weekly for 1 year.]

This stock went tidal on news that Diana, which still derives most of its revenue from distributing meat and fish, had turned itself into a provider of supposedly cutting-edge data switching devices.

But then the frenzied buying was over, and within months---hear the Dick Dale surf guitar?---wipeout. The stock has been treading water between the low 20s and mid 40s ever since.

Even at those prices, though, Diana is still nothing but a pipe dream according to New York investment advisors and highly visible short-sellers Asensio & Company. An active participant on several stock boards in THE MOTLEY FOOL, company president Manuel Asensio has issued a number of press releases denouncing Diana as nothing but a fraud run by scam artists. The stock isn't worth more than $5 a share, he says.

Asensio is hardly an unbiased source of information. With an undisclosed but presumably still significant short position in Diana, he has an interest in spreading negative news that will drive the stock price down. Some would even say he's willing to make up negative news. And many individuals who've crossed his path find him pompous, self-promotional, and generally annoying.

On the other hand, it's still relatively unusual for short-sellers to launch public campaigns against a company the way Asensio has. While investors may not like short-sellers, carefully considering the shorts' arguments, whatever their merits, can often be a constructive exercise. For one thing, it often becomes easier to dismiss a company's critics once you know what they really think. For another, shorts may open one's eyes to an ill-considered investment.

Last Tuesday, October 15th, Diana stockholders were encouraged to question what they do and don't understand about the company's prospects when Asensio & Company issued a strong sell recommendation in the form of what appeared to be a devastating 16-page report on the company's much-touted Internet technology. Asensio put out a press release summarizing the findings and took the unusual step of posting the report on the Web (available at www.asensio.com).

After conducting "an exhaustive technical software and hardware evaluation" of the company's products, Asensio concluded that Diana's "performance claims, which form the basis for any anticipated growth, are wholly false and incorrect." The company's products are "completely uncompetitive" and "it is technically and economically impossible to attempt their resurrection."

The report concluded that "the Company's performance will continue to deteriorate for the foreseeable future. We see no reason for any investor, particularly investors with a fiduciary duty, to purchase or hold this security."

Asensio's report hit the wires in the morning, before the market opened. Within the first hour, Diana's shares were trending down from $32 to a low of $28.75. The company then asked the New York Stock Exchange to halt trading in the issue pending a news release. Five hours later, the wire services picked up Diana's response to Asensio.

Without specifically addressing any of Asensio's charges, James J. Fiedler, Chairman and CEO of Sattel Communications (the Diana subisidiary that sells the switching technology) came out swinging.

"[Asensio's] purported 'news release' is false and misleading. This apparent manipulative style of 'news' releases is obviously only aimed at attempting to put pressure on the stock price by an acknowledged short seller. I believe that our customers who have employed our products would be greatly surprised to learn that features and capabilities that they use every day somehow do not exist, as alleged by the short seller."

The company's press release also indicated that Sattel was "in possession of a letter which states in part as follows: 'This letter provides written documentation regarding a request offered to me by a Manuel Asensio, representing Asensio & Co., New York, New York in September of this year, to perform a negatively biased engineering analysis on the Sattel products and organization for a fee of $5,000 dollars plus a percentage of the profits made short-selling Diana Corp. stock...' The analyst refused, citing the untruthfulness of the premise."

Finally, Fiedler said that Sattel has recently hired 12 new salespeople, additional R&D personnel, and a new CFO, suggesting the company was indeed growing. In a phone interview with ROGUE, Fiedler said that Sattel now has 20 fulltime employees working on R&D with 8-10 fulltime consultants at any point in time plus four other companies doing subcontract work. All total, he estimated R&D man-hours per month were in the high 40s. "We're probably spending 25-30% of our revenue on R&D," he said.

In the press release, Fiedler also said that Sattel expects to report about $4 million in revenues with a modest profit (instead of a forecasted loss) for the quarter ended September 30, 1996. When the stock re-opened about ten minutes before the close, it quickly rallied to end up $3.25 to $35.75.

Diana short-sellers were dismayed by these developments. They contended that Asensio had contracted an expert consultant with special insight into the relevant technical issues. In their view, Fiedler's response was an attempt to distract investors from the substance of Asensio's charges by refocusing attention on Asensio's inarguably self-serving motives.

Several shorts charged that Diana misrepresented the material nature of the news release in order to stop the bleeding in the stock. Asensio simply said that "the company's representations to the New York Stock Exchange were false and fraudulent." The shorts argued that the company was simply trying to buy time to work up a response and line up buyers for the stock once it re-opened.

"Absolutely false," said Fiedler. "I can tell you that people called us and asked what's going on, and we said we're preparing a response. Period. We didn't say what it was going to say. ... We didn't go out and solicit buyers for the stock."

Fiedler said he believed the halt was appropriate due to the "blatant falsehoods" in Asensio's report and the confusion in the market that it created. And as for the timing of the re-open, Fiedler said the company issued its press release at 2:19 pm, but that the NYSE wouldn't open the stock until the Dow Jones News Service picked it up, which took more than an hour. Fiedler said they probably would have just re-opened the stock the next day if they had known the trading would re-commence with only a few minutes left in the trading day.

For the last ten days, Asensio and Fiedler have been trading shots at each other. Asensio said his attorneys have served both Diana and Sattel with the demand that they produce the letter mentioned in their press release. In his own press release, Asensio affirmed that "Asensio & Company have never offered any cash, any fee, or any percentage of any trading profits to any analyst, advisor, or consultant for any negatively biased or untruthful report on the Diana Corp. or on any other company."

Fiedler declined to make the letter public, but in a follow-up news release this past Monday, Sattel indicated that they had received a second letter from a different consultant whom Asensio allegedly approached with a similar offer to produce a negative report on Sattel's technology. Fiedler said that the first letter came from an outside consultant who "has done work for us in the past" but that the second letter came from a consultant with no such ties to the company. Sattel has turned both letters over to their counsel, who are expected to pursue the matter with the proper authorities, presumably the SEC.

This public battle is great theater, but so much smoke does not necessarily guarantee a fire. On the other hand, Asensio is not the only party to express skepticism concerning Diana's prospects. Numerous reports in the financial media have puzzled over the company's surprising re-fashioning of itself from meat and fish distributor to Internet highflyer.

A Real Transformation?

Indeed, in a July article, ROGUE even examined claims made by some investors that Diana was an old-fashioned Wall Street promotion orchestrated by leading New York City hedge funds, some of which have recently been selling or registering to sell huge chunks of Diana (see the newly amended S3-A showing that Arthur and Joseph Samberg of New York's Dawson-Samberg, among others, plan to liquidate their holdings). The checkered histories of Diana CEO and chairman Richard Y. Fisher and president Donald E. Runge (who sold a large block of shares this summer) suggest that anything may be possible.

[Check out our earlier look at Diana: 7/23: When Should Investors Worry?.]

With 5.3 million fully diluted shares outstanding and a recent stock price of around $30 a share, Diana's market cap is $159 million. Yet the shorts make a strong case that neither current nor foreseeable earnings can justify this share price. The company lost $0.17 a share in 1995 and $0.76 in 1996. Trailing 12-month earnings show a loss of $0.93 a share. Asensio estimates that Diana will lose $1.25 a share for fiscal 1997, ending in March.

On September 4th, Diana announced its unaudited first quarter results for fiscal 1997. Net sales increased to $87.2 million from $81.5 in the first quarter of 1996. However, the company reported a net loss of $1.3 million ($0.27 per share based on shares oustanding prior to the recent 5% stock dividend) compared with a loss of $0.4 million ($0.10) in the first quarter of 1996.

The heavier losses followed from reduced earnings by Diana's C&L Communications business and increased losses at the Atlanta Provision Company, the meat and fish distribution unit that Diana has been unsuccessfully trying to sell off for more than a year. This unprofitable unit accounts for 85% of the company's sales.

The key to Diana's future, however, is supposed to be Sattel Communications, the now wholly-owned subsidiary that's in the business of marketing telecommunications switches. The unit was formed in 1994 as a joint venture with Sattel Technologies Incorporated (STI). At that point, Diana was trading at $3.50 a share.

In the last year, STI has swapped its entire 50% interest in Sattel for shares of Diana. Diana's recently amended 10K indicates that STI just cashed out its last 4% of Sattel in exchange for 15,000 shares of Diana. The latest amended S3-A filed Monday October 21st, however, shows that STI has registered to sell all of its 330,000 shares.

While Diana's stock is trading on great expectations for Sattel, switch sales for the last quarter were only $840,000, $796,000 coming from the sale of two Digital Switching Systems (DSS switches) to California-based Internet service provider Concentric Network Corporation (CNC) and the remaining $44,000 from what Fiedler described as a "small upgrade to a previous customer."

For the fiscal second quarter ending in September, Sattel's projected $4 million in revenues also come mainly from sales to Concentric. According to Fiedler, Sattel sold five DSS higher-capacity switches to Concentric in the second quarter, with each costing about $600 thousand, he said. The two switches sold in the first quarter were upgraded (at a cost of about $200 thousand each) in the first weeks of October and should appear as revenues in the third quarter.

Still, this means that of Sattel's $4.8 million in projected revenues for the first six months of fiscal 1997, a total of seven switches sold to Concentric accounted for $3.8 million in sales. Fiedler said the additional $1 million in sales for the second quarter came from shipments of of 2-3 switches to perhaps two other customers. He said he did not know yet what sales had been booked in the second quarter versus the third and could not release the names of customers at this point.

These numbers, however, suggest the importance of Sattel's contract with Concentric. In July, Diana announced that it had finalized a deal whereby Concentric would purchase 21 DSS switches within the next 12 months. According to Concentric spokesperson Donna Laughlin, CNC is in the process of transforming its 214 points-of-presence (POPS) in the U.S. and Canada into 21 SuperPOPS by the end of calendar year 1997, and each of these SuperPOPS will include a Sattel switch.

Originally, investors speculated that 21 switches would amount to $42 million in sales for Sattel. But the company's switches can be purchased in configurations costing anywhere from $125,000 to $4 million according to spokesperson Tony Squeglia. Customers typically buy the basic switch and then add capacity over an extended period of time as their business grows. In that sense, the ultimate value of the switch sales depend a great deal on the market success of Sattel's customers.

Fiedler says that the Concentric switches may generate $2 million a piece "over a long term horizon, but certainly not right away." If CNC reaches the success level they're shooting for, "yeah, they could be worth that much. But that's over a life cycle" of the switch, which could be several years. Though Fiedler suspects that Concentric may eventually purchase more switches as their business grows, he says "there is no commitment from Concentric for more than 21" switches at this time.

Based on Fiedler's comments, the sales of switches to Concentric seem to be on a pace to generate only about $12.6 million over the next year or so. Moreover, Diana last week amended its 10K filing with the SEC to indicate that the Memorandum of Understanding between Sattel and CNC was "nonbinding."

The Concentric deal has sparked widespread skepticism since it at least appears to have been secured by a $5 million bridge loan from Diana, suggesting that Diana has essentially paid Concentric to accept the DSS switches and thus validate the company's technology.

In a phone interview, Asensio said that he believes Concentric agreed to the switch deal because Concentric was unable to pay its overhead expenses at the time.

"[Concentric] could not make their payroll. There was a company that was doing the financing, the financing had failed, they were cash flow negative with no money in the bank. And they were offered $5 million in return for a very soft promise possibly to buy a telephone switch. Now that's a deal that anyone would make, and that's the deal they made."

Yet Fiedler says that these were "two very separate transactions."

"We were looking at putting up a wholesale network to sell to smaller, regional ISPs. We'd done the evaluation. We knew how much capital it was going to cost. Concentric made the suggestion to us that rather than build our own network and probably spend 12 months to do it in terms of putting up sites and so on that they were in the process of going out to do a round of funding, and that if we wanted a participate in the funding that a) we would get a stake in Concentric and secondly we would have access to the Concentric network at a very favorable cost. The advantage was that their network was already up and running."

Sattel orginally planned to invest about $5 million along with its recently announced strategic partner StreamLogic also pumping in $5 million. But CNC's funding was oversubscribed, meaning they only needed $2.5 million from Sattel. But because of delays in reaching an agreement over the wholesale business and delays cause by the involvement of international investors, Concentric's fund-raising deal didn't close in June as expected. At that point, CNC came to Fiedler and asked for a $5 million secured loan to tide them over. As security, CNC offered Sattel common stock, interest on the loan, and warrants equal to 15% of the stock.

In August 1996, the loan plus interest was converted into 3,729,110 shares of CNC Series D Preferred Stock. In September, Sattel sold 1,838,235 shares of this stock to StreamLogic for $2.5 million. As Fiedler explained, "The money that I had lent them was converted to equity. I got my stock, I got my warrants."

The recently amended 10K states emphatically that the $5 million loan "was not a condition of the Memorandum of Understanding" regarding the sale of switches. In addition, it appears that Sattel's investment in Concentric was limited to $2.5 million which has since been converted into what Fiedler estimated was a 2-3% equity position in Concentric.

"The deal was actually a good deal in my opinion for two aspects," Fiedler said. The distribution arrangement will allow Sattel to save perhaps 12 months and require less capital than would have been the case if Sattel had constructed its own distribution network. Second, Fiedler believes Concentric is a sound investment due to the company's expansion plans and content deals with WebTV, Intuit, and Sega Genesis. "I have faith in their strategy. Unlike lot of ISPs, they are really concentrating on content."

But confusion over the switch sales and loan to Concentric justifiably raised questions about Sattel's prospects. And the fact that CNC still accounts for the overwhelming majority of Sattel's sales should leave investors somewhat uneasy.

This is especially true since Diana is working with Hambrecht & Quist on a restructuring that, according to Diana's Fisher, would "make Sattel Communications a 'pure play' by separating it from Diana's other subsidiaries."

Fiedler said, "We absolutely could miss by a week or two, but we're shooting for early November" for announcing the details of the restructuring." It probably wouldn't be effective for another 45 or 60 days" since the documents would need to pass SEC scrutiny and the deal would need to be approved by shareholders. According to public documents, however, there "can be no assurances that any such transaction will take place, nor can there be any assurances as to the timing or value of any such transaction."

Since Diana's Sattel subsidiary is clearly supposed to be the company's principal growth vehicle going forward, shorts and other critics have focused on that part of the business. Last week's report represents Asensio's most thorough attempt yet to discredit Sattel's technology and burst the Diana bubble once and for all. The report argues that, rather than offering any performance advantage over competitors, as Diana has claimed, Sattel's technology is dated and all but obsolete with essentially no sales potential and certainly nothing that would justify Diana's current market cap.

In this past Monday's press release, Fiedler said that the report clearly indicates that Asensio "has no understanding of our products, technologies, competitors, and markets, and that his decision to purposefully communicate inaccurate and misleading information does a grave disservice to our customers, our shareholders, and our employees. . . . [W]e suggest that Mr. Asensio ask the consultant that wrote the report for a refund."

[Check out a description of Asensio's report: 10/25: Asensio's Report and Sattel's detailed response: 10/25: Sattel's Response.]

If Asensio contends that Sattel's technology simply cannot do what the company says it does, Fiedler argues that Asensio simply doesn't understand what the technology is supposed to do, or the problems it is meant to solve. One obvious way of figuring out if the technology works would be to talk with Sattel's customers. But Concentric has refused to go into any detail about the switches, and discussions with other Sattel customers have offered a mixed picture.

The inter-exchange carrier Lightcom has operations in San Antonio, Texas and main offices in Washington, DC. Lightcom bought two Sattel DSS 3000 switches about a year ago and has helped beta test the company's recently announced adjunct processor platform that provides enhanced 800 number translations and time of day routing features, among others.

According to Julian Bart, director of operations for Lightcom, the Sattel switches have worked well. There were some software bugs with the adjunct processor platform early on that have been fixed by Sattel. Overall, he said the company was satisfied with the product. He also said Lightcom can and might expand the switches up to 10,000 subscriber lines and that the company plans to buy three more switches over the next five years. He would not reveal what Lightcom paid for the switches.

Bart is also an enthusiastic supporter of DataNet. He said he's discussed the product at length with Sattel's developer and engineer. "To me it's probably one of the most innovative products to come down the pike." He said Lightcom doesn't need it yet. "But there's no doubt about it, we will."

He also said Lightcom doesn't have SS7 capability at the moment but he expects they will at some point. "We have not ordered it yet. We're waiting until we get a little bit more network up and then we'll go with the SS7."

Datacom (DCI), an ANSI and local exchange carrier that provides tandeming and blocks of minutes primarily for corporate customers in Miami, purchased a DSS switch in September. (Fiedler could not confirm whether this purchase landed into the second quarter or not.) The company has four employees and reported sales last year of less than half a million dollars. The switch is believed to cost about half that much and was secured with a capital lease.

Company president Anthony Harper said, "We're very satisfied with the switch. We're very satisfied with the company's support, as well." He said the switch includes SS7 capabilities. Though he said that it is not a distributed switch, he also said the switch is easily updated because of built-in redundancies that allow changes to be made "while it's hot."

Harper said he chose the Sattel switch for price and because it was more scaleable and versatile. He said software upgrades can be expensive but that Sattel offered an attractive deal in this regard.

A third Sattel customer contacted was Jerry Hamilton, director of international technical operations of Rochester-based Frontier Communications. Hamilton purchased a DSS96 in December of 1995 to link up to a switch in Russia, an application that was very small and static and thus required simply the lowest cost solution. The switch cost less than $100,000.

The switch has "been fine. It definitely has its place," he said. But Hamilton suggested this the switch is most useful for smaller applications.

"Once I get above where I'm at now, it stops becoming effective for me because when I start getting into the larger applications, then the other manufacturers of true international gateway switching (Northern Telecom, Erickson, etc)" start looking more attractive. "You go ahead with high up-front costs because the feature sets offered with the switch offset trying to grow it on the smaller basis."

He also said that Sattel doesn't currently offer the kinds of international SS7 variants that he needs, so it's unlikely he'll be in the market for more switches from them. "A prime consideration is SS7, and I have to talk on several different flavors of SS7. I want to be able to accomodate other types of signaling as well. I can do that as long as it's not SS7. The SS7 on the Sattel switch has not been deployed, it's behind."

"I need a great deal of flexibility without having to wait," he said. "As I roll out new features and services, I have to have a device that I know will interface with the rest of my network. So I'm going to tend to go with the vendor who is already known to provide those applications."

He also said that while "you can just keep tying the frames together" to get up to 10,000 ports, "that's not necessarily a real efficient way to do it because you lose some of the capacity by tying it to the frame next to it."

Fiedler said that Hamilton is correct. "He's talking about C7, the international version of SS7. C7 has an international version almost for every country." Sattel doesn't have a Russian variant of C7 running, which would make it impossible for Frontier to use Sattel technology with its Russian customer.

Despite Sattel's thin customer base at the moment, Fiedler believes the company is poised for tremendous growth. "We're sure as hell not going to stay at $4 million" per quarter, he said. "We obviously believe that we're on a growth curve, or I wouldn't be here."

He said the that the $4.8 million in revenues expected for the first half of fiscal 1997 "isn't going to be a fifth of the revenue" for the year. "Are we going to do more than double next year? Well, if we do our job."

Asked if he could justify the $200 target price bandied about earlier this year by Hambrecht & Quist analyst Joe Noel, Fiedler made an attempt. Taking the second quarter run rate of $4 million, and considering that the company made a little money, "it looks like we have a lot of capability in terms of earnings per share."

"I leave it to the gods as to what PE ratio they put on this, or any company by the way." But he said a PE of 40-50 based on his expectations for earnings would land the stock in Noel's range. For fiscal 1997, Fiedler said he expected $25 million revenue "or more." For fiscal 1998, "I would be grossly disappointed if we doubled it. A good company on a roll should be able to triple it." He said the margins are running 25%, suggesting the company could do $4 or $5 dollars a share, or total earnings of about $20 -25 million for fiscal 1998.

"So if you did a hundred million in revenue, you'd be there, based upon the way they're doing PE. ... Will I step up and guarantee that? Hell no. Do I think that's outrageous? No."

Given such optimistic predictions from Sattel's CEO, it's somewhat difficult to account for the actions of STI, the company that sold Diana the switching technology. Though there are several steps in STI's agreements with Diana, STI appears to have sold its stake in the once jointly owned subsidiary for 330,000 shares of Diana, currently valued at around $10 million. What's even more interesting, the recently filed amended S3 shows STI registering all of these shares for sale.

ROGUE could not reach George M. Weischadle, Sattel board member and chairman, president, CEO and majority stockholder of STI. So it's hard to know when STI will actually sell these shares. But it's at least curious that STI seems anxious to unload its stake sooner rather than later.

According to Asensio, "It's up to market participants to correctly price capital assets. The people who provide research, who are in the market, who give advice to people, those are the people who set the prices for stocks. It's a fallacy to sit back and think that the NYSE or the SEC is going to police the market to such an extent that frauds like this will not be perpetrated on the market."

Of course, it's not at all clear that Diana is a fraud. If an investor believes Fiedler's comments and the rebuttals issued by Sattel in response to Asensio, there seems to be nothing in Asensio's charges that should give investors pause. On the other hand, there's very little earnings visibility for Sattel despite Fiedler's ambitious plans for growth. Discussions with Sattel's few customers offer a mixed picture.

But more speech ought to be better than less when it comes to making markets more efficient. Whatever the merits or faults of Asensio's report, publishing it on the Web at least gives all investors access to the shorts' main arguments even as it clearly serves Asensio's own financial interests. Still, those arguments should make Diana investors look more closely at whether they do indeed understand the business they've invested in.

At the very least, the report has triggered a series of detailed financial and technical disclosures from Diana and Sattel that clarify the Sattel's business. But the soap opera, no doubt, will continue.

-- Louis Corrigan (RgeSeymour), 10/25/96

                          DIANA CORP

        Prices Available: 01/02/87 through 10/24/96

                                  Ticker: DNA

  Date       Volume     High/Ask     Low/Bid      Close
--------   ----------  ----------  ----------  ----------
10/27/95      45,682     10 5/16      9 1/2       9 3/4  
11/03/95      86,104     11 7/16      9 1/2      10 7/16 
11/10/95     359,681     13 3/8      11 1/16     12 11/16
11/17/95     138,734     12 13/16    11 5/8      12 11/16
11/24/95     102,630     13 9/16     12 3/4      13 5/16 
12/01/95     173,155     13 13/16    11 3/4      13 1/8  
12/08/95     614,945     21 1/8      12 11/16    20 1/8  
12/15/95     796,524     23 5/16     17 3/4      22 3/16 
12/22/95     297,049     24 13/16    20 9/16     24 13/16
12/29/95     147,683     25 9/16     23 1/4      25 5/16 
01/05/96     710,735     25 1/16     18 1/2      18 7/8  
01/12/96   1,552,945     19 1/2      12 1/4      15.20   
01/19/96   1,049,682     14 5/8      11 3/4      12 1/2  
01/26/96     409,682     16 3/8      12 9/16     14 1/2  
02/02/96     232,946     15 5/16     14 1/4      14 1/2  
02/09/96     159,155     15 1/16     13 11/16    14 15/16
02/16/96     167,999     16 1/16     14          14 1/8  
02/23/96     120,735     14 15/16    13 3/16     14 1/8  
03/01/96     920,735     21 1/8      14 1/4      19 3/8  
03/08/96     778,840     19.95       17 5/16     18 5/8  
03/15/96     397,787     21 1/2      18 1/2      20 3/16 
03/22/96     794,419     28 7/8      20 3/16     28 3/8  
03/29/96   1,143,471     29          25 7/16     27 5/16 
04/04/96     462,313     29 3/16     26 3/4      28 3/4  
04/12/96   1,575,472     38 13/16    27 13/16    37 3/4  
04/19/96   3,380,103     43 1/4      33 1/4      39 11/16
04/26/96   2,114,208     51 7/16     41 5/16     46 13/16
05/03/96     652,523     48.45       43.70       44 3/16 
05/10/96   3,250,734     74 15/16    43 9/16     63.65   
05/17/96   1,594,313     85 1/2      60 9/16     80 3/4  
05/24/96   1,804,313    114          79 15/16    98 5/16 
05/31/96   1,523,262    108.30       88 15/16    91 9/16 
06/07/96   3,020,420     90 15/16    62.70       67 3/16 
06/14/96   1,676,104     84 7/16     66 3/8      75 7/16 
06/21/96   2,053,682     77 13/16    51 9/16     55.10   
06/28/96   3,066,103     57 13/16    34 11/16    38 5/8  
07/05/96   1,501,682     47 7/8      35 1/2      40 1/4  
07/12/96   1,298,841     43.70       37 3/16     39 7/16 
07/19/96   1,473,472     39 3/16     23 3/4      24 1/4  
07/26/96     896,840     26 3/8      19 3/8      21 3/8  
08/02/96     910,734     26          20 11/16    22 9/16 
08/09/96     548,523     24.70       21 5/8      22 15/16
08/16/96     373,892     23 1/2      21 1/4      21 3/8  
08/23/96   2,862,840     33 1/4      19 1/2      29      
08/30/96     976,103     32 1/16     25 5/16     29 9/16 
09/06/96   1,453,473     39.90       28 1/4      39 11/16
09/13/96   2,206,287     48 1/2      38.95       45 1/4  
09/20/96   1,234,800     45 5/8      38          39 3/4  
09/27/96     646,300     42 7/8      38 1/4      41 1/2  
10/04/96     670,900     43 7/8      35 1/8      36 1/2  
10/11/96     608,300     38 1/8      32 3/8      33      
10/18/96     756,300     36 3/4      28 5/8      32 1/2  

ASENSIO'S REPORT

According to the Asensio report, "Sattel claims that its Digital Switching System ('DSS') telephone switch is 'technologically advanced' and offers features and performance capabilities that are either not available or superior to those of their competitors. The Company claims that the DSS can handle up to 375,000 Busy Hour Call Attempts (BHCA), which is equal to approximately 104 cells switched per second.

"It also claims that the DSS has an extraordinary expansion capability ('scalability') from 96 line side ports to a maximum line side port capacity of 10,240 ports. Each single 'line side port' services one subscriber's telephone line. In addition, Sattel claims Signal System Seven ('SS7') and ISDN connectivity. These claims form the basis for Sattel's large sales projections, which the Company has made but never attained.

"Even if Sattel's products actually performed according to the Company's claimed specifications the DSS would not in any way be unique or special. However, we believe the truth is that none of these claims are correct. In fact, we believe that because of inherent limitations in the DSS software and hardware technology it is not economically and technologically feasible for the DSS to ever achieve or even come close to achieving these performance claims, which are already being offered by many of its competitors. To the contrary, based on a complete hardware and software engineering evaluation, we believe that these products are inferior to those offered by the majority of Sattel's competitors."

The report says that the DSS was developed by Sattel Technology before 1989 as part of a large system sold to Indonesia's public telephone company to link hard-line telephone facilities to a Hughes satellite. The only significant sales of the product have occurred in Indonesia, where a total of 37 switches were sold prior to 1994.

In total, only 60 DSS switches have ever been sold. And Diana has made no significant improvements to the technology. Indeed, the report offers a detailed step-by-step analysis of the product's shortcomings.

In terms of hardware, the report argues that DSS uses a centralized switch matrix rather than a matrix distributed over line cards like more modern systems. This old design limits both the number of ports and the kinds of new features that can be added to the system, making it less competitive than rival systems designed by Summar Four or Excel.

Also, the line cards used by DSS have no processing capabilities, meaning the main processor must control every facet of the line cards' activities through the use of a dedicated control bus rather than an internal signaling packet such as a High Level Data Link Control.

"Such designs are rare, if not unheard of today, as virtually all modern switching designs employ distributed processing elements on line cards with dedicated internal packet bus to service them."

What this means, to take only one example, is that when an end user picks up a telephone handset, the dial tone must come from the central computer rather than the DSS line card. Smart line cards are essential if a system is to be readily extended.

The report also questions Sattel's claims for the system's call switching capacity. The faster switching models available now use the SS7 digitial protocol, and they can only handle 35 calls per second. DSS does not employ this advanced technology yet Sattel claims that DSS can handle three times as many calls.

"We believe Sattel's call capacity claim is false. Based on the capabilities of other more advanced platforms, and on the limitation inherent in the [sic] Sattel's DSS architecture, it is highly unlikely for the DSS to achieve this call handling performance."

The Asensio study also challenges Sattel's claim that DSS can support up to 10,240 ports. The dated DSS line cards can only support 16 ports per card while the DSS only has room for 42 DSS line cards, suggesting a total capacity of only 672 ports.

"It is physically impossible to support the line-side port capacity as specified by Sattel on the DSS platform in a single DSS enclosure utilizing their 16 port analog line card." And linking up several enclosures is "impractical if not impossible," and would make the system more costly than that of competitors.

The report raises other technical points as well. Unlike modern switches, the DSS depends on software loaded onto the Electrically Programmable Read Only Memory (EPROM) used for common household appliances, such as microwaves. This makes the system incredibly inflexible even though the dynamic market it targets requires "constant software updates and enhancements to make it compatible with a wide variety of telecommunications environments." This problem suggests "further deficiencies in the software operating system and dated software architecture as a whole."

In fact, this fundamental DSS architecture makes maintenance of the system difficult. Unlike modern products that use a graphical user interface on a PC with Windows and allow maintenance from essentially any terminal on a LAN or WAN, the DSS is maintained though Ascii terminals using characters only.

The lack of hard drive storage makes it impossible to download data from an ethernet port or store network management packages from third party vendors.

These structural problems follow in part from the dated "assembly language" software code that DSS employs. More modern switches use C, C++, or Java. Because of the technical expertise required to write new code for this unwieldy system, the study asserts that "Sattel will find the maintenance and evolution of its software near impossible." Even if that basic software were completely chunked, the system's fundamental architecture would still make the platform "unmaintainable."

These issues become a concern, in particular, when addressing the R1/R2 in-band signaling protocols, which actually vary in different foreign markets. Sattel says that DSS supports the generic international standards, but the DSS cannot be readily tailored to fit the specific protocols in each market since data tables cannot be downloaded into the switch (as Excel does). Nor can customized software from floppy disks be loaded onto a hard drive (as Summa Four does).

The R1/R2 protocols are embedded in the core EPROM software. Sattel will have to customize this generic code for each new country it goes into. Given the company's small engineering staff, this would appear to be a burdensome if not impossible task.

Finally, the report asserts that the DSS does not in fact possess SS7 or ISDN connectivity capabilities, important requirements for most advanced telecommunications networks. Sattel appears to be working to "patch" SS7 functions onto the platform by employing New-Net's AccessMANAGER product. "Most of Sattel's product claims actually belong to New-Net," the report says. But such system integration takes time, even for companies such as Summa Four which, unlike DSS, have a switch with a programmable interface.

Sattel is trying to build a custom interface that would run at 19,200 baud. But New-Net is designed to interface via a faster port, like an ethernet, which DSS does not support. So the limitations of the DSS will limit handling capacity regardless. "We believe that even if ultimately successful with [the] SS7 project, due to the limitations brought on by the switch matrix, interface, and line-card, the DSS will still in no way improve its competitive position."

The Asensio study also looks at Diana's claims for Sattel's DataNet switch, a switch that in theory could be used either separately or in conjunction with the DSS. DataNet is said to afford Internet Service Providers (ISPs) a significant cost advantage over other switching technology.

But Asensio says that DataNet is made of "off-the-shelf modems purchased from US Robotics" and that DataNet can only work in conjunction with a series of other third party technologies. The arcane specifics of this process should encourage investors to study the technology before leaping to invest in it.

While ISPs do need an efficient way to bring data from remote dial-up facilities to their central computer center, which is what Sattel insists it's offering, "many local calling solutions already exist, which are far more advanced than what Sattel is offering," the report says. Intelligent routers sold by Cisco, 3Com, Bay Networks, and Cabletron, "do far more than Sattel's DataNet, allowing conversion of the analog modem traffic into TCP-IP before even transporting over leased facilities."

Sattel's DataNet provides no such advantages, merely converting analog data to a digital format at one end of an ISP.

"Further, nothing technically significant is introduced at the ISP POP side that provides any advantage over other products now available from some of the above mentioned suppliers. DataNet is a very low level data multiplexing solution that any network manager could put together himself with a few days and a Black Box catalog. In short, DataNet realizes no new technical achievements or product abilities that the market has not long ago solved and has developed beyond."

The Asensio study concludes by examining Sattel's chances for selling switches to regional Bell operating companies, competitive local exchange carriers (such as MFS/Worldcom), wireless local loops, companies involved in shared tenant services, long-distance carriers, foreign companies, Internet Service Providers, or video server systems. In each case, the study concludes that Sattel's technology doesn't stand a chance.

In the Internet market, "better products already exist." And for most of the telco markets, the lack of SS7 or Advanced Intelligent Network protocols mean that DSS simply cannot provide what companies need. The basic hardware and software architecture will only add to the company's problems in marketing DSS overseas due to the R1/R2 protocol obstacles.

The study concludes that long distance, where even low end switches could be used, is the only viable market for DSS, but only if the product is "extremely price competitive." Strong competition from the likes of Summa Four, Excel, Harris, Redcom, and ITT/CoreTelco mean that the DSS switch "would be the bottom end product offering at best."

The report also argues that due to technological deficiencies, neither the DSS platform nor the DataNet product offer "anything that will give a customer any advantage, nor will it allow that customer to integrate services that they may require in the future from products that exist today, or will be developed."

SATTEL RESPONDS

In Monday's press release, Sattel's Fiedler charged that the Asensio report is "blatantly inaccurate and a gross misrepresentation and distortion of the facts. Our products use current technologies. Our DSS Switches are SS7 capable for central office (Class 5) switching operations; have a hierarchy of mass storage devices; have Ethernet capabilities; utilize X windows---a Unix-based graphical user interface; use distributed switching matrices in addition to its centralized matrix, and are scalable from 96 to 10,240 lines in multiple frames."

Fiedler said that contrary to Asensio's report, multiple frames are commonly used by major central office manufacturers such as Lucent, Nortel and Siemens. The idea that this is less efficient is "totally bogus," he said. And he said that the switches' distributed matrix uses line cards with Motorola single-chip processors that can handle sophisticated functions.

Indeed, in our phone interview, Fiedler expressed bewilderment at Asensio's charges regarding the switches' SS7 capabilities. "It runs on an adjunct processor, on a [Sun Microsystems] Sparc that you can run externally or plug into your chassis. What you do is you write the interface software for your switch and the set, you don't write SS7." According to Fiedler, other major switch manufacturers such as Nortel rely on NewNet's SS7 stack, which he says has over 50% market share.

In terms of maintenance Fiedler claims the Sattel switches have even more flexibility than products produced by Summa Fourand Excel, which are in any case not Class 5 switches and thus do not qualify under the new telecommunications law for use by alternative telephone service providers.

Calling him "Nonsensio," Fiedler laughed at Asensio's charge that the DSS switches lacked flexibility because they were coded in assembly language. "I got a real kick out of that cause years ago I use to write operating systems. When they're talking about only appliances use it, you should tell that to the guys at IBM and Sun Microsystems," Fiedler said. "All systems have to have some assembler, or they won't work."

"The core switching component of the DSS absolutely is in assembly language. The DSS also uses C++." He said that the switch functions are coded in assembly, but that all of the support functions such as the applications and the support for peripherals are in C+ or C++.

As for the company's DataNet capability, Fiedler reiterated the company's view that the product works and that it can lower an ISP's costs significantly as it allows them to transmit four times the amount of data that can be transmitted today over standard T1 lines.

"Let's put real blinders on and say, I'm not going to do anything but be an ISP, and number two, I don't care what my line costs are. Yeah, functionally, you can buy different kinds of equipment" such as that offered by Ascend, Cascade, or US Robotics.

"But forget about the equipment: the real problem is the line costs. Right now, you can go to any of the big ISPs today that do residential, or home-office kinds of customers. All of the ISPs are in the red at the gross margin level because of the flat-rate pricing."

In some cases, the line costs per customer run as high as $800 a month. But ISPs can only put 20 customers on a port. If they are only charging $19.95 per customer per month, they can't help but lose money since the would need to charge $40 per customer based on their basic line costs.

"That's the kind of thing we're replacing," Fiedler said. "It has nothing at all to do with, is the [competitors'] equipment good. I think the equipment is probably very good, but I can't afford $40 a month per person."

Fiedler said the key to lowering an ISP's costs is not the switches per se, but whether you can get rid of the line costs. Cheaper equipment won't help reduce costs nearly as much as cutting out these line charges. The Datanet allows ISPs to "get rid of a lot of the local access lines. We're co-located in the long-distance carrier's point of presence. And because we are a class 5 switch, we qualify under a lot of different tariffs to provide lower cost lines that do the same function.".

Sattel has said that they have customers who are using DSS switches with SS7 or Datanet installed. Yet in phone interviews with both Fiedler and Squeglia, neither would provide the names of these customers.

According to Fiedler, Concentric is not running the SS7 since it doesn't pay off at this point for Internet providers. The supposition all along has been that Concentric is indeed using the DataNet product. Even the amended 10K states that the non-binding Memorandum of Understanding signed with Concentric in May arranged for Sattel to "supply its DataNet product" to CNC 21 new SuperPOP sites.

However, ROGUE has not been able to confirm that this is the case. The amended S3 filing of October 21st details the limited sales history of the product. "Sattel did not have any sales of the DataNet product for the twelve months ended March 30, 1996 and the first quarter of fiscal 1997." But it indicates that Sattel has provided Concentric "with an option to purchase Datanet capability in the future."

Talks with Fiedler and Squeglia left the issue unsettled and Concentric's Laughlin would not detail the switch specifications, indicating she could only act as a reference for Sattel's comments. The presumption remains, however, that Concentric has purchased the DataNet product, at least for the latest round of switches to be delivered.


Main
Page


Fools &
Their Money


Hall of
Portfolios


Investing
Foolishly


Daily
News


Stock
Research


Foolish
Games

Copyright©1996, The Motley Fool
All Rights Reserved.
This material is for personal use only.
Republication and redissemination, including posting to news groups,
is expressly prohibited without the prior written consent of The Motley Fool, Inc.