December 2008 Archives

The advantages of unions

| 1 Comment
As I've mentioned before, we have a very major budget crunch facing the State of Washington. As WWU gets a good chunk of its funding from the State, this affects us as well. We've been told to expect an overall general fund budget shortfall of around 13% if I'm remembering right. This is a lot.

A week or so ago the Governor released her budget that she'll be submitting to the Legislature. It is guaranteed that there will be differences (and very likely major differences) between what she submits and what she'll ultimately sign. This budget does not include any new taxes.

As a brief aside, Washington has a referendum process where citizen-proposed initiatives can be put before the people for vote. In the past, any new taxes passed by the Legislature have almost always been thrown in front of the Initiative bus and run over. The State, therefore, is very gun-shy about tax increases. It is not surprising that the Governor is passing onto the Legislature the responsibility of suggesting that perhaps raising revenue is a way to get out of this mess.

However, the budget proposal does not include funding for contracted pay increases for certain unionized worker classes. Unsurprisingly, the unions are starting to sue for breach-of-contract. It's a bit puzzling that they're doing it now, when the proposal is just a proposal and no power beyond the power of suggestion, but they're doing it anyway.

This can lead to an interesting possible side-effect in the next 2-4 years as this budget crisis unfolds. If the unionized workers get their contracted increases and us salaried professionals don't get our annual cost-of-living increase, we could (er, will) see cases where the promotion for a classified job into an unclassified job results in a pay-cut. In the last 2 years I had a sizable salary increase because the classified workers underneath me on the org-chart were given a sizable pay increase by the Legislature. I got the increase in order to prevent a salary inversion. If the classified workers get their contracted increases and we get the 0% us salaried folk are expecting, it could happen again; whether or not we'll get any money remains to be seen.

There is a down-side to unions, though. They negotiate anywhere from yearly to every four years depending on the union. A deal cut when revenue projections are still negative will have scanty to zero increases, which will still be in place should State revenue suddenly increase. In this way Union pay for public works is even more of a lagging indicator of economic woes than State budgets already are.

Snow day

For a wonder, WWU is actually closed today. WWU is famous for being one of the only schools to be open when everything else is closed. They closed last year right after thanksgiving break because the twelve inches of snow we received Saturday and Sunday closed Bellingham up tight. People in the office were marvelling at that, since the last snow-day was in the 90's.

At the time, I heard the reason WWU stays open is that such a large percentage of our students and faculty live within walking distance of campus, so dangerous road conditions aren't the concern it is for your average K-12. In fact, last Friday was another instance of everyone but WWU being closed. I'm not sure what happened over the weekend to push us over into snow-day.

Myself? I've been snowbound since Wednesday. The hilly neighborhood I'm in has very bad hills, and are impassible without tire-chains. In fact, the number of SUVs at the bottom of either hill (the third is so obviously evil no one tries it when weather gets bad) is testament to the need for some kind of traction device, not just 4wd.

Since Tuesday night, I have had 15 inches of snow fall on my driveway. And more is falling right now. What we've had is a very cold mass of air parked overhead, as we get the usual stream of moisture. So we've received a heck of a lot of snow that would otherwise have fallen as hard rain. If the air were warmer, the amount of water falling from the sky would be normal for this time of year. But due to the cold air, it's falling as snow.

Yeah. Snow day today.

BrainShare, what next

As promised, some additional information.

The BrainShare Forum announcement of the cancellation. I've met Mr. Groneman. He works with the BrainShare crew, as does a lot of Novell, as he has to get the Support and SysOp areas set up. He's also the guy who is (I believe) in charge of the whole thing. He states clearly that this is not the end of BrainShare for good, this year's step was taken due to the apparently marked decrease in registrations they've received to date.

Brainstorming some replacement ideas for the short term. One item mentioned a lot in the above thread was to have an ATT Live event during the same week as BrainShare is now, presumably to allow the people with nonrefundable air tickets to get SOMETHING out of the travel.

GWAVACon is offering discounts to BrainShare attendees, now that their travel budgets suddenly have more room. They claim they have more than just GWAVA and GroupWise there.

There are some hard economic factors that figure in here, I'm sure. Until this year, I understand that BrainShare attendance had been growing. From personal experience, the groups that had seen the most growth were the IDM and Linux portions of the conference. More than once I've heard BrainShare attendees say it feels more like a Linux conference than anything else, any more. Still, they were seeing regular growth. To see a sharp decline? That's clearly overall economics at work.

I know that the Sponsors for BrainShare expect to reach a certain number of attendees, and if Novell can't deliver the head-count, it could go poorly. Also, the $1695 reg-fee is heavily subsidized by both Novell and sponsors. While direct per-person costs such as food-service are rather low, the sunk costs of the Salt Palace and other such items are a fixed cost, so having fewer people around means the apportioned fixed-cost-per-head is much higher when attendance is down a lot. A lot of those contracts were probably signed months ago and have no back-out clause, or those that do have punative penalties associated, so Novell isn't saving a lot of money by doing this.

Also, Novell is not the only group cancelling a big trade show. MacWorld will be Jobs-free this next time around, and there are others. I guess travel-bans are one of the first things that employers are putting in place in light of reduced income, and that's really taking a bite out of the convention-business.

I wonder how the big Las Vegas hotels will make out?

Novell has canceled BrainShare 2009

| 1 Comment

I didn't find out about it until this morning, as I was snowed in yesterday and wasn't paying as much attention as I should. But still. I'm one of those people who cited travel-budget shortages for not coming. This is major news!

This begs the question, "is BrainShare dead, or just this year?"

It could be either, but from what I'm hearing in the Novell Forums it sounds like there are still plans to put on a BrainShare 2010, should conditions improve. I can't go to BrainShare 2010, those aforementioned economic concerns ($5.6Bn budget shortfall in the state budget for fiscal 09-11), so it's kind of moot for me. But still.

There is a lot of stuff going on over in the forums. As they're web-based these days, and the interface doesn't suck, I might be able to link to interesting bits. In another post.

Still. duuuuuuuuuuuude.

Bad weather

We're due for a storm up here. On Monday, the forecast discussion from the National Weather Service said that all of the computer models were unusual agreement about the Fri-Sun environment. The jet stream had dipped out in the Gulf of Alaska, and that meant we were on the express lane for an arctic low.

The description is one I'm familiar with since I grew up in the American Midwest. This is what's called an Alberta Clipper, since in the Midwest the weather pattern described above has the jet-stream dip over the Rockies instead, which meant the weather systems came out of Alberta. Only, this system is, you know, 1500 miles West of where it should be.

Whatcom County is in a special weather enclave itself. If you go to the Google Maps page of Blaine, WA and click Terrain, you can see what I mean. Bellingham is at the south end of the Fraser River valley. That river valley acts as a corridor for cold wind when the weather is right, and it'll be right this weekend. When the low passes over the Cascades, it'll park over eastern Washington and the wind direction will change. This is where the wind will start roaring down the Fraser.

The forecast for the next week is unusually cold, with highs expected to be in the 20's. If we get much snow before that, it'll be a bad thing with all these hilly streets; we'll get a lot of iced up streets. On the good side it should be sunny, so road that gets sun should melt.

The one time I had a friend with a GPS at my house we figured it was around 650 feet. A bit lower than I expected, since we usually got snow when the snow-level dropped to1000ft. There must be something in the microclimate around my house, because we seem to have a virtual elevation of between 900-1200ft. Since the only way to the country road from my house involves three very steep, and shady, hills, it is entirely possible I'll be snowed in until Monday. We'll see.

The price of storage

| 1 Comment
I've had cause to do the math lately, which I'll spare you :). But as of the best numbers I have, the cost of 1GB of space on the EVA6100 is about $16.22. Probably more, since this 6100 was created out of the carcass of an EVA3000, and I don't know what percentage of parts from the old 3000 are still in the 6100 and thus can't apportion the costs right.

For the EVA4400, which we have filled with FATA drives, the cost is $3.03.

Suddenly, the case for Dynamic Storage Technology (formerly known as Shadow Volumes) in OES can be made in economic terms. Yowza.

The above numbers do not include backup rotation costs. Those costs can vary from $3/GB to $15/GB depending on what you're doing with the data in the backup rotation.

Why is the cost of the EVA6100 so much greater than the EVA4400?
  1. The EVA6100 uses 10K RPM 300GB FibreChannel disks, where the the EVA4400 uses 7.2K RPM 1TB (or is it 450GB?) FATA drives. The cost-per-gig on FC is vastly higher than it is on fibre-ATA.
  2. Most of the drives in the EVA6100 were purchased back when 300GB FC drives cost over $2000 each.
  3. The EVA6100 controller and cabinets just plain cost more than the EVA4400, because it can expand farther.
To put it into a bit of perspective, lets take the example of a 1TB volume of, "unorganized file data", the seemingly official term for "file-server". If you place that 1TB of data on the EVA6100, that data consumes $16609.28 worth of storage. So what if 70% of that data hasn't been modified in a year (not unreasonable), and is then put on the EVA4400 instead? So you'd have 307GB on the 6100 and 717GB on the 4400. Your storage cost now drops to $5909.75. That's real money.

When you can't trust tcpdump

I just spent a good chunk of today bothering the telecom people to try and figure out why one server of mine couldn't talk to any off campus NTP servers. I had two servers, one was talking just fine, the other wasn't. For proof, I had packet traces showing that the non-working server was not getting an appropriate "ntp server" return packet.

And yet, after the telecom people sniffed the border firewall connections they saw UDP/123 packets on both sides. In other words, it transited the firewall just peachy. And also our IDS/IPS. So, clearly it was getting in. But it wasn't showing up on the tcpdump output on the server.

Then about 15 minutes ago I turned off the "restrict default ignore" line on that server.

And it started syncing off campus just fine. With packets.

WHY was tcpdump not showing the packets? That's what I want to know! Somehow, the UDP packets were being dropped before tcpdump saw them. Strange.

Infiltrating the market

Over on The Open Road there is a very interesting blog post. It talks about how Microsoft and Red Hat approach the market, and touches on Novell.
Microsoft offers a full ecosystem of software to would-be buyers, but its greatest success may actually result from its strategy to present customers with an "and" decision when initially purchasing Microsoft technology, rather than a difficult "or" decision.
And I really see this. The argument has been made internally that what you get from a Microsoft Enterprise CAL is worlds above what we can get from a Novell academic seat license, which follows into cost-effective discussions (and not good ones). It is soooooo easy to go all Microsoft, whereas a pure Linux solution requires a lot of stitching, and translation-glue.

The article goes on to point out that Red Hat's targeting people looking to do forklift upgrades from Unix to Linux. And then points out that Microsoft wins more of that traffic than Red Hat does, by a good margin. Largely because the Microsoft family of products is very complete.

As it points out, Novell figured this out a few years ago when they launched their collaboration with Microsoft. The fruits of which arrived today with OES2 SP1, and the Novell CIFS stack and Domain Services for Windows. This allows OES2 to do something you can't do with Samba (yet), pretend to be a full up AD Domain Controller.

And yeah, Novell's current marketing slogan is, "Making IT work as One," which is a clear embracing of the "and" concept described. If they could make DSfW work on plain SLES, it may make it an even more attractive product for people.

OES2 SP1 ships!

The NetWare 7 that never was

| 1 Comment
My last post generated some comments lamenting where NetWare has gone. I hear ya.

I have friends and have spoken with people at BrainShare who were closer to things than I was regarding how the next version of NetWare evolved. And to be truthful, it sounded a lot like how Microsoft moved from XP to Vista. If you'll recall, "the version of Windows after XP," was something of a moving target for many years. I recall media reports of Microsoft scrapping the whole project and starting afresh at least once.

My very first BrainShare was 2001, and that was the release party for NetWare 6. It was in 2003 when Novell bought Ximian, and bought SuSE, so it is clear when Novell probably decided to bet the house on this whole Linux thing. Yet at BS01 there was talk about NW7, or if there would be a NW6.1 version out. The rumors I remember from back then had NW7 being a progression towards a more application-friendly environment. I also remember hearing the L word around once or twice.

What we actually got was NetWare 6.5, which solidified NetWare 6 and made the core services better and more mature. What it wasn't was any more application friendly than NetWare 6 was (or even NetWare 5.1 for that matter). NetWare 6.5 released in August of 2003, the same month as the Ximian purchase. This is what tells me that Novell had decided on a path for NetWare 7, and it was green, not red. Open Enterprise Server arrived in 2005, which gives OES a solid year and a half dev-time between when SuSE was bought and when we started seeing public betas of OES. The NetWare version of OES was NetWare 6.5 SP3.

What happened to NetWare 7? It got lost on the roadmap. When NW6 came out, Novell probably had 6.5 on the roadmap as the next rev, with NW7 next down. The rumors we were hearing were very provisional, as the spot on the map held by NW7 was at least 3 years away. Sometime between BrainShare 2001 and when Novell started buying its way into the Linux world NW7 was dropped and the decision was made to port to a completely different Kernel. That decision was probably made in the summer of 2003, as the NetWare 6.5 development was entering final beta, and the task of allocating developer resources for the next full rev needed to be made.

Which brings us to today. OES2 SP1 is going to drop any day now, probably in time for Novell's quarterly earnings report. SP1 finally brings the Linux-kernel 'NetWare Services' to feature-comparable with the NetWare kernel. There are still a few things missing, like an eDirectory integrated SLP server, but now all the major points are covered. If you count it up, this has taken Novell a bit over 5 years to get to this point.

In my opinion, that's about right for an organization the size of Novell. Porting over the proprietary NetWare services to completely new kernel requires a LOT of developer attention, and Novell is a lot smaller than Microsoft. Also of note, it took Microsoft 5 years to give us Vista after XP, including the presumed nuke-and-rewrite they did. Novell got a boost in that they had already ported eDirectory to Linux, so that helped out the NCP side. But that didn't help the NSS folks, who had to figure out a way to do a NetWare-style rich metadata file-system on a kernel and driver model that expects POSIX-spartan file-systems. The problems Novell had with this were amply displayed in the performance problems reported with OES1-FCS. Samba doesn't scale to the same levels as CIFS-on-NetWare did, so that meant Novell had to create their own CIFS stack from scratch. The AFP stack on Linux is a joke, and the resurgence of Apple since 2003 meant they had to do something about that as well; by making a proprietary AFP stack. All of this represents nuke-and-rebuild-from-spec, which takes time.

Novell probably should have started the migration in 2000 instead of 2003. They already knew that Exchange 5.5 upgrades were driving a LOT of customers into Active Directory, which was triggering migrations away from NetWare. But, there are business concerns here. Novell managed to survive the fall of NetWare by diversifying their product portfolio enough that GroupWise, Zen, and Identity Management could support the company. It took until this year to return to the black, but they did it. Had they shot the NetWare cash cow two years earlier, it is entirely possible that Novell couldn't have survived the lean years.