Modern Middle Manager
Primarily my musings on the practical application of technology and management principles at a financial services company.
Still Working the Citrix Metaframe Angle

Friday, May 30, 2003  

We recently put another long-unused feature of our Cisco Catalyst 6509 to use -- to create a Citrix Metaframe server farm using server load balancing in Cisco's IOS. Any Catalyst 6509 with the MSFC card can do this. The upside is that we didn't have to pay extra for the Metaframe XPa or XPe licenses to perform the load balancing, instead using cheaper XPs licenses and hiding the servers behind the switch. The trick is to make sure that the clients and the servers are on different subnets. After that, it's pretty easy. Below is our configuration:

! Create a NAT pool for clients -- simplifies routing.
ip slb natpool CITRIX-CLIENTS 10.1.3.1 10.1.3.254 netmask 255.255.255.0
!
! Line up the server farm. This declares which real servers will be part of the farm.
! Connections are distributed by a weighted least-connections algorithm after the first
! four connections are made to the server. The maximum number of connections allowed
! is 20. Note one service is set for "no inservice". We had to take it out of the farm
! because it was misbehaving.
!
ip slb serverfarm CITRIXFARM
nat server
predictor leastconns
nat client CITRIX-CLIENTS
!
real 172.31.144.71
weight 4
maxconns 20
inservice
!
real 172.31.144.72
weight 4
maxconns 20
inservice
!
real 172.31.144.73
weight 4
maxconns 20
no inservice
!
! Create the virtual server. We used a subnet completely different from the clients & servers.
!
ip slb vserver CITRIX
virtual 172.31.145.79 tcp 0
serverfarm CITRIXFARM
inservice
!
! The following command is necessary to get slb working.
!
mls flow ip full

The cool part is that by melding server load balancing with a blade server we can scale our farm very quickly. Got to remember to test each new server before unleashing it on the end-users, though...

posted by Henry Jenkins | 5/30/2003 06:33:00 PM
(0) comments
Desktop Support  

We've reduced our desktop support calls 50% year over year. I attribute that to the desktop support position I was able to fill, dedicating a single person to help desk issues. Using his insight we've been able to solve many of the recurring issues and still keep support calls down while upgrading to Windows 2000 and migrating half our company to thin client computing. I'm very pleased, as are our end-users.

The next initiative I have is the ability to work the help desk remotely. I see a day when any of the IS staff can whip open a laptop, connect to the network via the wireless (ubiquitous and accelerated) Internet and identify/repair/manage every node on the network. The only part not falling into place is the wireless Internet. Hotspots aren't nearly as available as they should be and the service itself is expensive. And once my staff learns how to use it, it's only a matter of time before even sales figures out how to make it work...

posted by Henry Jenkins | 5/30/2003 01:01:00 PM
(0) comments
The Asshat School of Management  

We'll call this theory of management ASM for short. ASM is predicated on the following principles:

1. Information is shared on a "need to know" basis, and you don't need to know.
2. Risk is to be avoided at all costs. Risk brings humiliation and potential termination.
3. Most decisions are made by senior management; major decisions made by subordinates are routinely overridden.
4. Lack of cooperation between departments at high levels.

These four points guarantee slow decision making, fiefdoms and competition between lines of business. It also guarantees a loss of motivation, loss of initiative and loss of creativity. Employees are not motivated by keeping them in the dark and telling them they can't make good decisions. Only by opening up the decision making process can employees understand HOW to make better decisions. God forbid we create a learning type of organization!

What is amazing is that these practices are reflected in another milieu -- Arab military practices (see Walter Russell Mead's article, "Why Arabs Lose Wars" for reference). And we know how soundly those armies are defeated by those who use information strategically. We are sitting ducks, the Libya of financial services. Extending the metaphor, a couple of cruise missiles from an enemy and we're pretty much done.

posted by Henry Jenkins | 5/30/2003 12:32:00 PM
(0) comments
Reviewing My Job  

Since I'm obviously in a proper frame of mind to evaluate my position here, let's look at the positives and negatives, the advantages and disadvantages, the virtues & vices of being CTO at my current place of work:

The Good
-------------------------------------------------
Flexible schedule.
Pretty reasonable department discretion with a large (to me) budget.
Good staff that acts as a team, receives mentoring and continues sporadically to become better professionals.
Final authority on technology -- can implement OSS projects to solve a problem if I think it best, for example.
Direct vendor contact (helps with the perks).
Rest of company is pretty easy to work with, except senior management (see below).


The Bad
-------------------------------------------------
No mentoring.
No strategic direction for the company.
No motivation other than a paycheck and near-sinecure.
Stupid, stupid requests.
Staff needs occasional beatings.


The Ugly
-------------------------------------------------
Senior management appears to adhere to the Asshat School of Management (thank you Oliver Willis for the term).
Management at our parent company appears to do the same.

posted by Henry Jenkins | 5/30/2003 12:11:00 PM
(0) comments
My Job is Like Watching Television  

It's both mind-numbing and time-wasting -- simultaneously!

Today's work has been a melange of absurdity and aggravation. To wit: my two biggest tasks due next week are writing a new vendor management policy (we need one? Why don't you write them up -- our compliance officer is too busy to do, you know, compliance type stuff) and printing out a report for our sister investment management company. In the meantime I spend about 15 minutes convincing my systems engineer that maybe, just maybe, he should look into whatever the problem is with the website that made it go down. Wouldn't that be wonderful? Yes! Yes, it would!

It's a "blinking colon" kind of day, as I watch the hour and minute separator on my phone count each second out. Blinking colon sounds like a physical condition.

My temptation today is to think of new initiatives to chase after, new goals, new ideas. We're not done with our existing projects, however, so piling on work would be stupid and trying to implement initiatives on my own is a waste of my staff's ability and a demotivator. I think I'm rationalizing taking the rest of the day off and reading the entire Day by Day Cartoon series, then launching into Slashdot's archives. Or read something by Camus...it seems appropriate.

Maybe I'll write that newsletter column that I was tapped for. Did I mention my boss had a tizzy when I got approval to do it? Seems I didn't tell him first and he hates that...jeezus, management here is damn near spastic. Must be their blinking colons.

posted by Henry Jenkins | 5/30/2003 11:42:00 AM
(0) comments
I'm Not "The Computer Guy"

Thursday, May 22, 2003  

I recently confused the Sweetie over dinner with her friends when I acted a little cagey about my position. While part of it was due to crankiness, part of it is also due to the overwhelming banality of responses I get when I tell people what my position is. A typical conversation starts:

Them: "So, what do you do?"
Me: "I'm the Chief Technology Officer for a small financial company in Orange County."
Them: "Oh, you work with computers."

At this point I usually hear a selection from the following genres:

Patronizing: "Oh, you picked a great field. Computers will be around for a long time."
Opportunistic: "You know, I was thinking of getting a laptop for home. What's good?"
Clueless: "Hey, every time I try chatting with my friends on AOL and attach a picture it makes my computer reboot. Any idea why that would happen?"

I oversee the careers of five people, control a budget of around $1.6 million a year, set the strategy for the information services department and touch every aspect of the organization. Oh yeah, and some of my job involves computers. Maybe it's because I've worked my way up through the ranks, maybe because of all the hard work in every aspect I've cited above. I've got pride in what I've accomplished and as much as I don't expect a CFO to be asked about how cool Excel is or why a home calculator doesn't work quite right if you hold down the 0 and 4 keys simultaneously, neither do I desire nor require brain-dead comments about my profession. So there. And get a Dell, dude.

posted by Henry Jenkins | 5/22/2003 10:45:00 PM
(0) comments
Can't Sleep Yet

Wednesday, May 21, 2003  

No matter how sleep deprived I get, there are times when I just can't seem to fall asleep and my mind starts wandering. Tonight I wax nostalgic, thinking of authors whose books I've savored over the years, authors who will no longer spin their imagination into print. I've always been a fan of fantasy and science-fiction writers, especially those with an off-beat sense, whether of humor or reality. It seems like only yesterday that Roger Zelazny and Damon Knight died. Zelazny's passing is the most unfortunate for me; although his Amber novels were beginning to caricature themselves, without him I would be unable to claim that my favorite heroes were smartasses.

posted by Henry Jenkins | 5/21/2003 12:02:00 AM
(0) comments
The Lakers are a Great Example...

Tuesday, May 20, 2003  

...of an inconsistent team. This year they proved once again that talent doesn't mean that they're the best. You've got the talented workaholic who tries to do it all (Kobe), the extremely talented but stupid and self-centered guy (Shaq) and the roleplayers who have a specific purpose and are expected to simply perform consistently. Shaq wants to be The Man but refuses to lead by example. Kobe realizes that he's currently the #2 guy and can become the #1 guy when Shaq wears out his welcome. He appears to be frustrated that others on the team don't put in his time and effort to be the best. The other guys are drawing a paycheck and basically want to win, but there's only so far that talent and hard work can take them. How often does an IT staff resemble this situation? And how does good ole Phil get them functional again?

Consistency can be the hardest thing to achieve in team-building. I can get great things out of my staff but it's uneven. Sometimes part of what they do is great and part is so bad it undoes the greatness done. When poor results are achieved because of sloppiness rather than ignorance, what then is the solution? Laying down the law isn't the issue -- it's figuring out what the motivating factor is behind the inconsistent performance and addressing it. I've got such a meeting today. Do I use the iron fist? Or the velvet glove? A combination? What will be the carrot(s) and/or stick(s) necessary this time? I'm a problem-solving kind of guy and people are the biggest puzzle I've ever encountered...

posted by Henry Jenkins | 5/20/2003 12:14:00 AM
(0) comments
Dirt Doesn't Need Luck

Monday, May 19, 2003  

OK, I don't know what that means, either.

Moments from today:

Coming in at 10am because one of my staff forgot what the words "quality control" mean in English. Tomorrow he gets to learn.

Picking a fight with the Banking Services Manager because his damn staff keeps trying to install instant messaging software but our security only lets it half-install, thus (*Q#@&$^(#*&^$ up their systems and calling the help desk for support.

Meeting with the Real Estate department because they bought into the thin client strategy wholeheartedly and I rewarded them with a Harry and David's basket that sent them into a blissful sugary coma. Mmm, moose munch...

Waiting 15 minutes for a status meeting with my boss -- he didn't show.

Called a guy at the appointed time who's heading the new data center building project to try to answer a question he had raised last week and subsequently forgotten.

I need sleep.

posted by Henry Jenkins | 5/19/2003 11:34:00 PM
(0) comments
Project Kagome

Friday, May 16, 2003  

One of the more idiosyncratic practices we have is to name our Citrix Metaframe servers after women in sci-fi. Recently that's narrowed to female characters in anime. The latest Citrix project is named Project Kagome from Inu-Yasha. The name has no real significance other than as an ode to our goofiness; however, the project has been an interesting stretch for us. The overall project goals were:

1. Provide near-desktop performance for most applications.
2. Provide high availability.
3. Decrease instability and "quirks" in the user environment.

A number of choices led us to the Citrix environment for many of our desktop users. From that starting point we implemented Citrix Metaframe XPs on Vmware's ESX Server and got to about 22 users before performance problems finally caught us. That initial implementation has allowed us to really clamp down on quirks and bugs to the point of extremely stable sessions. However, we're not coming through on issue #1 -- performance. So how are we going to deal with that while preserving goals 2 & 3?

Our plan was to use a Dell PowerEdge 1655MC with three blades to serve as the Metaframe servers and use our core switch, a Cisco Catalyst 6509 to perform server load balancing (SLB). SLB is key here -- by publishing a virtual server that routes to a farm of real servers, we can keep the number of Citrix users per server down while also giving us the ability to take one out of service for work. In practice, we have three blades -- one is the baseline/test server, the others are the deployed servers doing time in the farm. SLB will balance between these two production blades and any additional ones we may add later. Finally, we are documenting each step of the way to make the process easily duplicated. With Dell's Remote Install software we're able to image the server at each step of the way. Today we performed some internal quality testing and beta deployment should start on Monday. We could be in full production by the end of next week. It looks like this will fulfill goals 1 through 3 while also giving us the scalability and disaster recovery capabilities we need. As for how many Citrix users each server will handle, well...we'll find out next week.

posted by Henry Jenkins | 5/16/2003 07:15:00 PM
(0) comments
Meeting with CL

Wednesday, May 14, 2003  

Today I met with CL to go over three items: to see if she was set up appropriately at her new office (she was), to demonstrate the new imaging system (she loved it) and to discuss upgrading her office to Windows 2K and Office 2K. When I proposed Citrix as a possible solution she both rolled and crossed her eyes in a manner I thought might not be reversible. Fortunately for her it was. The conversation was very similar to the one I had role-played with my girlfriend last night (she played the role of CL, very convincingly I might add). After this display I knew it was sales pitch time. The conversation went something like this:

CL: (rolls eyes)
Me: Of course, we would never roll out a solution unless we were delivering a perfect product. The quirks, bugs and performance issues we discover in our trials at this office will serve in creating that kind of product. If you're still not comfortable when we're ready, we can discuss other alternatives.
CL: I'm sure you would never give us a substandard product. I know how seriously your department takes our needs.
Me: Absolutely we do! Most of my conversations were with the prior department manager; do you know why I'm interested in rolling out Citrix to all of the offices?
CL: Money.
Me: Yes, that's the primary factor. If we save $50,000 a year in PC upgrades it's a decent sum of money. However, it's more than just that. Every dollar I spend is created from $15 in revenue so $50K a year is really $750,000 in revenue that we don't have to raise.
CL: Whatever the amount, it makes a difference. I'm willing to play with the team as long as we don't sacrifice performance for my group.
Me: That's my goal as well. In fact, we're about to put into production a system that should increase performance to a level you'll be comfortable with.


Besides the ass-kissing and my obvious solipsism with the revenue numbers, I've got her open to the possibility. I also learned that I'm not the only one pushing this in management and that if I can deliver what I'm promising (which I believe I can) then it should go through. Heh.

posted by Henry Jenkins | 5/14/2003 09:12:00 PM
(0) comments
Sometimes I'm Just Overwraught  

Having someone else look at your work can bring a new perspective. The last post? Pretty thick-headed. Very blunt. The goal: achieve my strategy by changing CL's mind, not bludgeoning her to death. The tune: what can *I* do for *her*. Let's give it a try tomorrow.

posted by Henry Jenkins | 5/14/2003 01:24:00 AM
(0) comments
Watch Your Back While Scratching Others

Tuesday, May 13, 2003  

Today I found out that I have a senior manager who is going bananas over the Citrix thin client strategy I'm trying to pursue. And I don't mean bananas in a good way, I mean bananas as in wild and psychotic. Baby. Bathwater. Tossed out. You get the idea. We'll call her Crazy Lady, CL for short. Always willing to make a big deal over small potatoes, I find CL is bad-mouthing the thin client strategy I'm trying to pursue and is, in fact, trying to roll it back. The last department CL tried it on rebelled in an easygoing way, letting her know that all was well and that they really liked the concept and the actual implementation. I see I'm going to have an uphill battle with this, as this manager tends to ignore all reason when she's in this sort of fugue. Unfortunately I don't have the budget to upgrade her 24 person office with new PC's when we roll out Windows 2K and Office 2K. So, it's time for a strategy:

1. Identify the person or people whose thin client stories are poisoning the well. I know two off the top of my head. There might be three or four overall. I can neutralize at least one with a good lunch discussion.

2. Start the public relations machine. I need more good stories told in e-mail and at senior management meetings. That means dropping in on the latest senior managers who have been converted and getting a good story told. Surveys through e-mail help as well because it's documented.

3. Get a smaller branch converted, shake out the bugs and let them sing the praises of the implementation. Much like the department that rebelled against CL, get them to love it. This means a lot of high touch, rewards for being the first branch, etc. Never underestimate the power of reciprocity. They will be treated like royalty until they're comfortable. This will give me two major departments that work similar to her group. Both will get handsome bribes for their willingness to cooperate. Seventeen-inch LCD monitors sound good, don't they? Sure they do! Thanks for playing.

4. With that initial track record and the necessity of upgrading older PC's at CL's office, they will have some choices to make -- either run the Windows 2000/Office 2000 tandem on older machines with the concomitant speed decrease and disruption of work, or convert to thin client and see it work seamlessly. The easiest way to sell it is to start with a lunch meeting describing the existing track record, showing the financial benefits of going thin client and then asking for a pilot program to prove the concept. If that doesn't work, once they've experienced the system slowdown from an upgrade they will want a change and we'll do it in a guerilla fashion.

Don't get me wrong -- there are very few times that I will work around a senior manager. However, there are also times I'm willing to risk to pursue a vision for the company that I think is right. I think this is right. And I think I can get it done with the velvet glove.

posted by Henry Jenkins | 5/13/2003 05:03:00 PM
(0) comments
Blade Servers, Part III  

OK, so a new error crops up. When I roll back to a prior image the blade server appears Unmanaged, which in Dellspeak means, "We can't image your server any more." After spending half a day doing what I thought was reinstalling the remote install agent, I found that I had to bypass the autorun.exe on the Dell Server Management CD, go straight to the directory that had the Windows remote install agents and reinstall them directly. Voila! It worked. Back to installing apps...

posted by Henry Jenkins | 5/13/2003 04:39:00 PM
(0) comments
My First Wireless Woes

Monday, May 12, 2003  

Suddenly my notebook running 802.11g stops working. I can't get a signal, yet I'm about 6 feet away from the access point without obstruction. It was the damndest thing. I finally tried changing the SSID and channel and, lo!, it worked. And another default SSID appeared in my choices. Apparently I have a neighbor with a wireless network that conflicts with my own. Well, sounds like someone needs a big ole DoS attack...haha. Just kidding. Really, though, defaults are bad and I should have changed that before. Once again, laziness strikes and the results are, well, frustrating. Laziness really doesn't hit hard because that requires effort. Right? Is this thing on?

posted by Henry Jenkins | 5/12/2003 11:19:00 PM
(0) comments
Finishing Up the Day...  

Meetings occupied about half my day today (no, nothing interesting), planning meetings a little more and a semi-crisis with our Metaframe strategy when the server crawled after last night's Acrobat installation. Well, that's why backups are supposed to be made. Good thing I listen to my own advice. Heh. It's a bit of a black eye, unfortunately, as my boss' rival is using it to bad-mouth the entire department. Sigh. You would think people would have better things to do...

The new strategy is to get a load-balanced set of XPs blade servers to handle the load (and the eventual crash -- this is Windows, after all). This should keep senior management off my back while reinvigorating the idea. Plus a whole lot of ass-kissing thanks to today's performance.

posted by Henry Jenkins | 5/12/2003 11:13:00 PM
(0) comments
Blade Servers, Part II  

After opening the boxes and going through the documentation, only about 20 pages were necessary. The first thing to set up is the Embedded Remote Access (ERA) module. This module turns the blades on or off, provides access to the Network Service Module (NSM) and probably does some other things I didn't wait to figure out. The NSM command language is similar to Cisco's IOS so I had a four gigabit etherchannel set up in about 5 minutes. I plugged in the KVM switch and found out that the PrintScreen key toggled between servers. Then I installed Dell's Systems Management CD to a server so I could use their management tools on the blade server. Finally, I was ready to go.

Some vendors, such as RLX, claim that software is installed to a "virtual" server first before being deployed to a blade. Well, Dell did things a little cheaply -- a blade has to be used to create the images. I broke out the USB CD-ROM that came with the system, installed it on the first blade and booted up with Dell's OpenManage ServerAssistant. Wonder of wonders, it worked. I installed Windows 2000 Server with ServerAssistant, rebooted, and got started on creating a Citrix Metaframe server.

The most obnoxious part was figuring out how to capture an image from a blade. The way to do it is:
1. Install the Remote Install agent from Dell's Systems Management CD.
2. Reboot the blade, setting the boot sequence so that PXE boot occurs before the hard disk.
3. Wait for the OS to boot up.
Once it boots up and the management server contacts the blade server successfully, imaging can be performed. The PXE boot threw me the first dozen times.

posted by Henry Jenkins | 5/12/2003 07:23:00 PM
(0) comments
It Is SO a Word!

Saturday, May 10, 2003  

The ever-so-contrary girlfriend argued with me that "incentivize" is not a word, it is management babblespeak. According to dictionary.com:

To offer incentives or an incentive to; motivate: “This bill will help incentivize everybody to solve that part of the problem” (Richard A. Gephardt).

Not only is it a word, sweetie, the example cited is by one of your own, a Democrat. Mwahahahahaha...

posted by Henry Jenkins | 5/10/2003 02:50:00 PM
(0) comments
Blade Servers, Part I

Thursday, May 08, 2003  

Although I've written about the concept of using blade servers in our environment before, I am happy to announce that I've finally gotten my hands on one for real -- a Dell PowerEdge 1655MC with three blades. Each blade is a (kinda wussy) dual PIII running at 1.4GHz with 2GB RAM and two mirrored 73GB Ultra160 drives. The impetus behind getting the 1655MC was to set up a Citrix Metaframe server farm. Our current Citrix servers are running on VMware's GSX Server and the interactive performance is just too laggy. I'm expecting that access to two processors, even if they're slower than the one used by the virtual servers, enhances performance. My next entry on this server will be: how do I set it up?

posted by Henry Jenkins | 5/08/2003 10:36:00 PM
(0) comments
Profits and Probability

Wednesday, May 07, 2003  

About 60% of our business is in the personal trust, ahem, I mean wealth management industry (I forgot that we're rebranding that department). As I've mentioned before, we take wealthy people's money and keep it from their kids (and the government if at all possible).

The trust business is unpredictable and tempermental. Most trusts are unique by nature; everybody has their own goals in the management of their assets while they are alive and after they're gone. Most wealthy people have a variety of assets: cash, stocks, bonds, real estate, annuities, businesses, limited partnerships and the like. In addition, trust documents are legally binding contracts. Add into the equation that the trustee (the company running the trust, aka my firm) is required by law to provide for the beneficiaries of the trust (i.e., those who get money from the trust right now) as well as take into consideration the "remaindermen" (those who get the assets divided up for them when the trust distributes). Like politicians and money, those factors create a volatile mix -- one that can easily end in litigation at some point along the way, either because of the nature of a particular asset (e.g., real estate) or because the beneficiaries and the remaindermen have conflicting agendas.

The ubiquity of litigation results in the need for risk management and the application of probability, put succinctly as, "What's our likelihood of getting sued if we take this piece of business?" Taking some risk is necessary if the company wants to make any money at all. Taking too much risk could put the company out of business with litigation damages. So how does the company determine what business to take? By applying certain controls and reviews, management attempts to weed out business that may result in litigation down the road. Most analyses are done by instinct, but let's use a more scientific approach. Let's say we take a piece of business expected to last 5 years resulting in $100,000 in fees. Although it seems like a slam-dunk trust, there is always some risk of litigation, say 5%, resulting in a maximum guesstimate (based on experience) of $200,000 in damages and legal fees. Would you take this business? The expected return on this trust would be $100K - (5% X $200K) or $90K. I would say yes. This is the application of probability to the analysis of new business. It's like the example of flipping a fair coin where heads results in you winning $200 and tails results in you winning $100. Would you take the bet? Statistically you should. The expected return is $200 X 50% minus $100 X 50% equalling $50. That's a winning bet.

Taking this thought further, risk management should be an integral part of the sales process with management review able to quantify, as best as possible, why new business should or should not be accepted. The sales force and senior management should be incentivized on the expected return, wouldn't you think? What if, instead, the sales force and senior management were given incentives based on revenue goals. Or even net income, unadjusted for risk. And what if, on top of that, a new position was created to try to clean up and manage the risky business brought in by that sales force after the contracts are signed. Would that bring in trust business that was more or less likely to be exposed to litigation? Would it create a sales force that attempted to bring in riskier types of business just to run up the company's revenues and their commission checks? No, my company couldn't possibly be that shortsighted.

posted by Henry Jenkins | 5/07/2003 12:26:00 PM
(0) comments
Word Mojo

Monday, May 05, 2003  

I'm hooked on this stupid game. It's like a scrabble/crossword puzzle hybrid. And it feeds off of obsessive-compulsive personalities...yes...

posted by Henry Jenkins | 5/05/2003 10:59:00 PM
(0) comments
Chapter 4  

Chapter 4 of the Executive's Guide to Information Technology describes their preference for a well-organized IT department. They believe in splitting the IT department into two parts -- Application Development and IT Operations. These are further broken down (e.g., IT Operations are broken down into help desk, end-user support, systems admin, network admin and telecom). The authors believe that the positions should be ranked so that mobility from one area to another serves as a career ladder. Although I haven't worked in larger organizations, the organization they map out sounds good to me.

Interestingly enough, my IT department is organized similar to their suggested best practice. I have five employees -- two are programmers and three are IT Operations. I have further broken down the IT Operations specializations into Desktop Support, Systems Engineer and Network Engineer. There are some compromises that must be made, of course. Several roles will overlap and there's hardly a method for me to promote the systems engineer into the network engineer position. For one of my top performers I am considering a cross-discipline position, something similar to a project leader, in which I plan to help him develop management skills. The issue of IT management is discussed in Chapter 5.

posted by Henry Jenkins | 5/05/2003 10:20:00 PM
(0) comments
I'm #9!  

Number nine for the search "strive for mediocrity" on Google! Bringing you tepid management and information technology thoughts for almost six months!

posted by Henry Jenkins | 5/05/2003 10:16:00 PM
(0) comments
More Server Consolidation  

The Experiment of the Month has been getting a Citrix Metaframe server running under VMware's ESX Server on a dual-Pentium 4 Xeon Dell PowerEdge 2650. The installation went flawlessly and the Citrix Metaframe server we used had been developed under VMware's GSX Server so it migrated smoothly to ESX. Two major issues cropped up once the server was in production: performance and backups.

The backup problem is relatively simple -- ESX Server looks like it's a modified Linux kernel and, as such, should be able to back up over the network. However, it uses a proprietary file system called VMFS that can't be accessed by BackupExec's backup agent for Linux. To back up the VMFS files, one has to use a tool called vmkfstools that converts the virtual disk to files that can be saved. At this point we're looking for a solution -- perhaps suspending the server in the wee hours of the morning with a Perl script, copying the VMFS files to the Linux file system and moving them to our filer.

The second (and more noticeable) problem was performance. The performance of Microsoft's Office suite was terrible with only 6-8 users on the system. Outlook took a while to load, Word and Excel would frequently pause. Interactive performance, overall, was ugly. Although we opened the network pipe from 100Mbps to 1Gbps, it still worked poorly. Finally we came across a new product, Expedian from Wyse. The software appears to tune memory requirements for certain DLL's so that they don't take forever to load. I'm not sure exactly what it's doing to "optimize" the system, but I now have 20 users on simultaneously without any interactive performance problems. This is an impressive product. I'm very pleased.

posted by Henry Jenkins | 5/05/2003 02:34:00 PM
(0) comments
Weekend Reading  

Few, if any, corporations will hand you promotions or awards strictly based on merit. Corporate politicking is what positions people to attain what they want. This weekend's reading was by James Carville and Paul Begala called
Buck Up, Suck Up... and Come Back When You Foul Up.
Carville and Begala, in a Southern, down-home, folksy manner that begs for a dope-slap, outline their twelve principles for success in politics. Most of these seem intuitively true. In my own professional career I can see where I've used several of these principles. I found their discussion of "How to Communicate", "Frame the Debate" and "Know What to do When You Win" to cast some interesting light on those subjects. This book could be some help as a starting point for those wondering how to play the game.

On the other hand, except for their successes with the Clintons, most of their political efforts have yielded a litany of failures that they've chronicled (to their credit) within this book. They believe it was their hard work and dedication that led them to the apex of success rather than any other explanation, such as tapping into the prevailing zeitgeist. As such, maybe the only truly concrete lesson worth taking from this that they got this book printed before they left another 20 years of political failure in their wake.

posted by Henry Jenkins | 5/05/2003 01:36:00 PM
(0) comments
search
the author
archives
links
open source
vendors
stats
reading