Slashdot Mirror
The 3Com Saga
prostoalex writes "A flashback to 5 years ago reveals 3Com as a global multi-billion dollar company, respected and revered around the world. Today Bob Metcalfe's creation is a money-losing $2 billion dollar operation trying to find its niche. The 3Com Saga from Network World magazine takes a look at the history of 3Com Corp."
Simple. People stopped using dialup and 3Com couldn't keep up in the NIC card market. But seriously, 3Com used to make some very high quality modems when I first started getting into the internet.
Are you being sarcastic? Most built-in ethernet controllers use 3com chipsets, and all that's done for me is make my life a LOT easier to deal with, because you only have to really worry about 3 chipsets after that: a builtin 3com, a builtin intel (though they seem to be appearing less and less frequently now) and maybe the tulip driver for netgears for 3-4 year old PCs.
Never shoulda given up Palm...remember when Palm's were all subranded US Robotics, and then 3Com?
Actually, I'm kinda talkin' outta my a$$ here...I have no idea if selling Palm was a good business decision or not. I just know I've always like PalmOS, and 3Com used to get some advertising every time I put my PDA on its cradle...
SO YOU'RE GOING TO DIE: The Comic for Dealing with Death
I also remember a time when Telus was in the process of dumping 3Com ADSL modems for DLink of all companies, because DLink made a better business case (read better price) for a product that was 99.9% as reliable as the 3Com product.
So to sum it all up I think that 3Com sort of let things get away by simply not keeping up with the economy of computing, not the raw technology. They still have all kinds of respect in the industry, they just have to re-learn how to sell themselves.
That really is my homepage, no kidding.
IMO would have to be Intel. I find Intel's NICs to be pretty rock solid, and pretty resonable in terms of their prices.
Not to mention that lots of mobos ship with Intel's NICs built right in.
Anyone with me on this?
Sunny Dubey
As much as I have to agree with you, I now regret my Linksys purchases. I should have gone with D-Link or Netgear.
I've had to stop using Linksys cards in all my Windows machines because the poor drivers would bluescreen *every* time I ran a particular class of network-intensive application (ie: P2P stuff like BitTorrent). FreeBSD is quite happy with the hardware (higher quality drivers?), but suffice to say that the quality of Linksys' own drivers is very, very poor.
And the little Linksys router I've had for a while locks up after running for a couple of days in full DHCP mode - it's fine with a manual IP configuration, but that's not exactly convenient so I'm about to replace it with a D-Link router if the latest BIOS update doesn't fix the problem.
Good point, but once they do they will cease to be 3Com, and would more accurately be called "Faceless Holding Group Company of No Discernable Character".
At that point you are right - anybody with a finance degree and a bit of capital can and will compete with them. And probably win.
3Com was a tech company once, but now it is losing its identity.
Almost anyone that is still using dialup is either
1) lacking any other options
2) doesn't give a flying monkey about the performance of their modem
Group 2 will be entirely happy with their motherboards soft-modem which negates a lot of the demand for real ones.
On that note, I'd like to add more of the blame to the rise of integrated networking on just about every motherboard these days. Whereas before, when I would always go with an Intel or 3Com NIC, these days the integrated NICs on most motherboards(specifically Nvidia nForce and Intel based boards) work almost just as well as a 3Com solution, complete with the low CPU overhead(the only loss is a couple of truely Pro features), but it's already there, and it's free.
I can hook up my integrated NIC to my well-priced Linksys router, and do today what 5 years ago would have been many times today's cost. 3Com caters in part to a market that really no longer exists.
That's a very damning look at 3com... customers are supposed to be one of the most important things to a company (because if they're not happy, what're they going to do? Not come back, that's what) but the article insinuates that they dumped them like a cart full of wrinkled potatoes.
I still like their products... or did like their products, I'm not sure how this outsourcing will affect their quality of goods. It seems like they're stripping away the one thing that people still like about them.
Oh well. I wish them luck.
----- Wtcher Dragon, UDIC
The differentiating aspect of tech firms is that many have huge cash hordes from the .com bubble that will sustain long past their expected expiration date. Sun for example will take at least a decade to die given their cash horde, notwithstanding their inability to generate profits.
Of course that idea died and few companies are splitting off their best tech at this point.
I disagree with your glowing commendation of everything LinkSys. These are the geniuses who's early model DSL routers would reveal the login password by clicking on "View Source" at the login page.
With my first LinkSys DSL router I found that Internet traffic would inevitably become sluggish or just stop working after several days of heavy use and would not behave normally until I cycled the power. When I mentioned this to a friend he told me that he had two LinkSys routers at his office plugged into lamp timers so they cycled power every night. I took it as a joke until I visited his office one day and saw the timers attached to the wall beside the routers!
I moved up to a newer LinkSys router to support a couple of laptops via wireless connection. The range was weak and connections would frequently drop when both laptops were used simultaneously.
I'm currently using a US Robotics (3COM) wireless DSL router that is giving me no problems at all.
I must admit, however, that Linksys has the first wireless DSL router I've seen with actual support for Wi-Fi Protected Access. I'm still waiting for a firmware upgrade on my USR box to support that.
"Why would you buy a network card when nearly every mainboard has one built-in? And even the chipsets are losing to Intel..."
Because the add-on can do things that the built-in can't e.g. hardware cryptography (VPN, RIAA, hint, hint).
Typically an onboard chip/network card will use more proccesor and memory resources then a regular add in card, while the performance will vary, I have noticed that 3com's 3c905 series will off load the most and will be the fastest with the least amount of extra resource draw.
Most people won't even think twice about using an onboard nic. I don't really blame them. but in a windows 2000 domain with active directory switching nics from a built in 3com to an add on 3com had users thinking they got a new workstation. The speed difference was especially noticed when using aplications that reside on the server and have to check back in everytime data is changed or needed.
Just somethign to think about when using the on board nics. If a little extra performance would be nice or needed.
One thing I have noticed about the low-end gateway routers is that the older products are almost always more reliable than the newer ones. I have an old Netgear RT314 that I'm sure still works fine, but I stopped using it because it didn't have an MTU limiting (MSS clamping) function. I got a Netgear MR814, and it dies when too many connections try to go through it. V1 Linksys BEFSR41s were pretty stable. V2 was more unstable, and V3 is completely useless. I have no historical impression of D-Link routers because they haven't worked well on the local DSL service until fairly recently.
I used up all my sick days, so I'm calling in dead.
The 3com Sega ?
I was trying to think of a cool switch that was 3 times as good as a Cisco but failed because Cisco said they'd have something better in a year or two...which ended up being as good but definitely not better...yeah..
There are many situations where I need more than one card. Onboard NICs, especially when they're Realtek, are terribly unreliable as well (the failure rate for onboard Realtek I've dealt with may be as high as 1 in 4). As much as you'd like onboard Intel, don't count on it in most boards when a manufacturer can save $3.
/do/ you care about?
3Com's problem? Go find their product on a big box store shelf. IF they have it (most won't carry it - it simply won't sell and shelf space commands a premium), it's $59-$79 per unit, compared to $16.95 to $19.95 a unit for Belkin/Dlink/Linksys. Often consumers can find a $5 to $15 rebate (combination of in-store and mail in) on top of that price.
According to 3Com's reps I've talked with lately, this is not viewed as a problem "because that's a market we don't care about anymore." Obviously, as you don't compete. But what
Filing patents. Honestly, check this link. It would appear to some that 3Com's heading down the path of the future dot-com: the tech litigator/patent holding company. (Good news: David Boies might have another client after SCO can no longer pay his rates)
3Com's been pushing firewall on a stick. Embedded firewall (via proprietary standard) technology, which has a little appeal but is too esoteric for the PHBs, and too lacking interoperability for the geeks. The current employment market makes it somewhat unwise for the recommender geek to push a hard to explain proprietary niche product to the boss when everyone else's product interoperates just fine (at a much lower price) without this feature.
The Chinese will steal'em blind. They can kiss their IPs goodbye. Watch'em lose their shirts in China...
ELOI, ELOI, LAMA SABACHTHANI!?
Cisco vs. 3com:
Recently I found out that our NBX call processor software has a DOS vulenerability. It wasn't a real big deal since the call processor isn't accessible via the internet, but since i'm a good network admin, I want to patch the box anyway. I go to 3com's website thinking I can download the updated code right? - Wrong. I had to call my "authorized reseller" and get the code from them. This just gave my reseller an excuse to come visit me and try to sell me more crap.
Cisco, on the other hand, makes you buy a yearly support contract (not too expensive), and they give you access to their TAC site. Login to TAC, download updated software, install updated software - done.
I'm a corporate customer and I like Cisco's method of support. I suspect alot of 3com customers feel the same.
3com are you listening? I don't want to call a salesperson every time I need to patch a box!
-ted
The joint venture with Huawei lets 3Com ride shotgun over product development while ridding its own payroll of expensive engineering talent and manufacturing plants.
"A Chinese engineer costs one-sixth of an American engineer..."
"A year ago 3Com had about 300 engineers on its payroll developing product. Today, we have closer to 900 engineers working on our behalf. Yet the cost of this is all off our books," Claflin says.
Nuf' said.
Their "Open" phone system that uses the H3 protocol which is published just where??? (rumor is at MIT -- anyone want to hunt for it?)
It also seems that all the version of the NBX software up to 4.1.21 have GPL code in them and you can prove it by just asking it. To upgrade to a version where 3com isn't illegally using GPLed code, you have to buy another license. Keep in mind that 3com was one of the few IT compaines that supported the DMCA. Maybe its because the DMCA helps hide the fact they are using open source software without following the terms of its licenses. Details are here.
One other nice thing about their new license scam is once your dealer goes under, your out of luck and when 3com can't find the prior license, you get to rebuy all them all over again. Too bad the best source for info on it NBX Group has given up on the product and is bailing out.
Applications like Outlook/Exchange?
I work for a federal gov't agency that purchased 30,000 Gateway workstations last year with these god-awful intel onboard nics. I can't prove the root cause, but we get dozens of calls for performance issues with applications like Arcview, web-based apps, etc. We're converting 45,000 users to Outlook/Exchange at the moment and the workstations just lock up for 10-30 seconds at a time during network operations, many times per day.
But how do I prove that it's the onboard NIC when the management won't even consider the possibility that they bought flaky hardware!!! I don't know how the hell to cure this one without just recommending that they add 2nd network cards but that as a recommended "fix" will get me in trouble because, according to the vendor, everything is fine and dandy with these systems. Makes me fucking mad. The Feds play the CYA game at master-level.
One thing to consider in this, though, are the newer crop of GigE cards that are integrated in the northbridge. Intel and Nvidia's newer chipsets are shipping with these sorts of chips, as they start to realize the same thing Apple did a while back: GigE plus anything disk intensive can and will max out a traditional 32-bit 33mhz PCI Bus. I dig 3Com products as much as the next guy, but I'll be damned if they could produce a worthwhile Northbridge assembly.
As a side note, their cards are also ridiculously expensive compared to the stuff that a lot of other manufacturers offer, and as has already been pointed out, the other stuff is sufficient for 98% of the jobs.
....that their biggest enemy was Intel.
They were right, as Intel went and built their own NICs into Intel-branded boards and took away a huge add-in market. The problem's even more obvuious in the notebook market where there are Centrino solutions from Intel, and other third-party wireless solutions - but nothing from 3Com.
Then again, who can blame them?
Now even the tethered Ethernet chipsets are too commoditized for Intel to be using their own in the D875PBZ series mobos.
Not that 3Com's been especially savvy or well-behaved about it all. They seem to have a bipolar problem... they would do great "engineering" products that were crap for the market they were intended (managed home office NICs anyone?) or they would completely miss segments they could do well in (such as HP did with their ProCurve switches) which came in at only a modest premium over 3Com's hub products when they were introduced.
As plenty of other posters have mentioned though - you get what you pay for.
Try and run an intensive app like eEye's Retina using built-in (or soft-NICs), and get ready to lose any other connection pending until the app has finished doing its thing.
To have ambition was my ambition.
I say they're ALL crap. Too quick to market with new features. We've owned three Linksys routers, one D-Link, and one NetGear.
:(
Linksys - One of the routers died after a month, one of them died after two years. The other is still working after 3 years. Not a very good success rate if you ask me, and I wasn't all that impressed with the feature set anyway.
D-Link - Quite a bit better than the Linksys crap, and the built-in printer server rocked - even worked with Linux (no driver required). However, one of the antennae fell off after a couple of week (no roughness, just fell off in my hand), and it's impossible to find replacement antennae. The speed of it wasn't so great when the firewall was enabled either, but it did it's job. The wireless range was attrocious! Oh yeah, it was only able to do one firmware upgrade and then all others after that failed
NetGear - Our current router is the Super-G WGT624 NetGear. I have been very pleased with it. Automatically checks for newer firmware, TONS of features that I haven't seen in any other router, very fast even with SPI enabled, easy to limit wireless by MAC, awesome interface, etc. We did have a couple of problems, however. When we first got it, the web interface would lock up regularly. It also was sometimes flakey about saving settings. However, both have seemed to stopped since I installed the last couple firmware updates. Now it is running problem free and I couldn't be happier.
I've heard a lot of complaints from NetGear, and we mustn't forget the time server fiasco, but I've been happy.
In summary, Linksys sucks all around, D-Link was useable but slow, somewhat featureless, and shoddy construction, and NetGear has been sweet for us.
In my experience, they started performing better than 3Com cards. Replacing 3Com 3C900B/C cards with Intel PRO/100S cards in my P133 server and PII-400 workstation, Linux 2.4.x, Unex switch) gave me a performance increase of roughly 2x (from 3-4MB/s to 7-8MB/s.)
I used to work for 3Com as a tech support bob; I was one of the people laid off in 2000 when they wimped out of the medium and big-end market.
It was a pretty strange day; we got in to work as normal, only to hear wild rumours about the US HQ EOL-ing entire families of projects - including some brand new ones like the Corebuilder 9000 (and some ancient ones like the Netbuilder II).
We were told that there was an announcement downstairs; at which point they laid out the whole gory mess to us; massive layoffs, company shrinkage and retreat-in-confusion from the majority of 3Com's market areas.
As tech support, we were given three months' notice with the possibility of moving to support small hubs and low-end kit (not exactly a challenging prospect!); what most of our lot did was get straight on the phone to ex-colleagues working for Cisco.
What really got us down over the next week or so was speaking to customers. There were multi-billion dollar customers who we knew well and got on with; and 3Com's sales people had been selling them new Corebuilders the week before; suddenly, they had no upgrade path, they'd invested millions of pounds uselessly, and they were not happy with us. A couple of the bigger companies were demanding that Eric Benhamou fly over in person and tell them exactly why 3Com had sold them something one week and discontinued it the next.
(The understanding customers at least told us that they didn't blame us for random acts of management; but it still wasn't fun for us; the company we were a part of was pissing all over our customer base, and we were at the front line as public-facing employees.)
There were days during my four years at 3Com when announcements went round that made my blood run cold. The USR merger was one; 3Com putting an NT server (on a blade) in a switch was another; but the biggest one by far was The Day We Wimped Out.
I think 3Com as a company deserves a graceful death. I still insist on their NICs (which are rock solid and have never given me a day's grief); but I have no need or desire for any of their other products.
The whole Huawei tie-up is another Bad Idea, IMHO. I've had to try and configure one of those things; the interface was terrible (it was just post-Cisco-lawsuit), the hardware was laughably unreliable (bad mainboard *and* two bad interface cards) and customer support (at least in the UK) was pretty much non-existant.
Eventually, we ripped it out and replaced it with a Cisco. If 3Com are relying on rebadged Huawei kit to recapture their share of the market, then I think they're on a hiding to nothing, and the 3Com name will take another nasty dent.
They had it; they blew it.
Gideon.
If 3Com fails, it will be mostly from self-inflicted wounds. Anyone out there who works with Total Control gear and worked with it back when it was USR Total Control can tell you how badly service and support fell off after it became 3Com Total Control following the buyout of USR. My former employer (an ISP) was a 100% TC shop, but we switched to Cisco RAS gear instead, and the greatest (by far) factor in that decision was 3Com attitude.
We could hardly get anyone at 3Com to pay us the time of day, we went over two years without seeing out sales rep, and that was despite being an active customer during that period. When we started to evaluate Cisco, Cisco was all over it. If we had an issue of any kind during the test period, our vendor would take it straight to Cisco if their own engineers didn't have an answer. At one point, we were having some problems that would be solved by the latest code for our modem cards, and the Cisco engineer who wrote it personally brought it to our office on a CD-R.
Naturally enough, when it came time to make a buying decision, we were unanimous in favor of Cisco. That was about three years ago, so I can't comment on whether 3Com has changed for the better or worse, or is still that way, but the worst thing that ever happened to USR, IMO, was becoming part of 3Com.
I like 3Com gear - Total Control is great stuff - and I have a couple of Superstack II switches right here in my den, and 3Com NICs in a lot of my computers, but they shot themselves in the foot through greed and poor customer service.
My father in law works for Unisys.
Unisys' real strength at this point is that they employ smart people at the consulting level. They don't really innovate. However, they are a big name in the business, and when you need a solution to a problem and you want it done yesterday, regardless of the cost, you call Unisys. They can move the technological earth for you.
Amusing anecdote.
A couple of years ago, when my not-yet-then father in law was working his way up the consulting chain, he ended up being the go-to guy for MBNA's head office. If you don't know who they are, someday look at where your credit card offers come from in the mail - chances are 1 of 3 comes from Wilmington, Deleware. That's MBNA - they are the elite of the super-rich credit card companies.
So anyway, MBNA used to use Sprint for their internet connection. And we're talking massive bandwidth - this is for credit card processing, so it's multiple fiber pipes, OC-3 size each, if not bigger - and they have to be up 5 nines percent or better. So, this is a multi-hundred-million dollar contract with Sprint.
So, one day, MBNA's connection goes down. And they're losing money, to the tune of something like ten thousand dollars a second. My father in law, who's name is Mike, is called in. They're on the phone with sprint, and nothing's happening - sprint promises to look into the problem and do line testing, and call them back within 24 hours. This is obviously unacceptable. Everyone's running around, mass chaos, cats sleeping with dogs, etc. The scenario ends with the C.E.O., the VP of something, and Mike, and a few more underlings of both MBNA and Unisys in a confrence room. The CEO has the VP call sprint, and work his way up to talking to the highest guy he can get ahold of in their tech department. He asks how long it will be, gets an answer, and hangs up the phone. He looks at the CEO and says, "They say it will be 12 hours".
The CEO looks at Mike, and says, "Switch it to AT&T."
Mike calls AT&T (through Unisys channels - and when Unisys calls, people listen). AT&T has locals there in 10 minutes, and people on planes from New York in 20. They move heaven and earth have it up and running in 2 hours.
Anyway, all that is to say, Unisys isn't dead - they've just shifted into a different market - being the power behind the consulting. From what I've seen, though, they (and a lot of other companies out there) tend to hire people, use them for a few years, and let them go before they get too many raises.
~Will
sig?
Hmm. What I read from your description of 3com onboard being wildly inferior to 3com on a network card matches my experience. 3com onboard is so *bad* it makes 3com on a card look wildly better by comparison. But if you need wildly better performance on a card, invest in a Kingston 4-port card: those still use the old DEC Tulip chipsets or a newer chipset with the DEC scraped off and a new manufacturer. These are not "tulip compatible", they're actually Tulip chipsets, and goodness do they work well under high load. Interrupt coalescing, here I come!