This time we'll dig a bit deeper, and see what kinds of information may be used to generate service targets (SVTs).

In a Help Desk situation you want to be aware of SLAs concerning resolution time/closure time and such - you also want to know how soon someone was assigned, and then started working the issue. SVTs can be set so they 'attach' to a ticket that follow a given START criteria (Service = Help Desk AND Status = Assigned - for example), and then do a STOP criteria (Service = Help Desk AND Status = Resolved/Closed). You get the idea.

Along the way you want milestones - say your goal is 24 hours for the case to be closed, and you want to be warned 12 hours into the ticket by email or a pager note - absolutely possible.

This was just a basic explanation of an SVT for a simple Help Desk ticket, but I am quite sure you can follow the enormous potential of such a product. Now, if you have the ITSM suite you'd get an integrated set of products that are aware of each other and ready to work with each other. There are templates you can apply and you could also create your own customized templates specific to your organization.

In order to create an SVT you first need a data source (DS). You could create your own DS based on an existing infrastructure (say an existing database of widgets that have status information in them, which you want to monitor). Or, you could use the integrated DS - such as Help Desk, Change Management, Compliance and such.

You could also use data from other sources (such as BMC Performance Manager: PATROL and the BPM Portal; SNMP Traps; Transaction Manager: TM-ART - in the case of BMC's (Remedy) SLM product) that you have set up in your organization and mine them for specific triggers. Say you want to monitor how long the CPU of an important machine stays at 60% or above for SLA purposes: sure, you not a problem. You can look at the history of each of these Measurement Records and figure out what data was retrieved and when.

It's easy to see the limitless applications of SLM - once you have a reliable, accurate data source you'll be able to consume its data, manipulate them and calculate all sorts of statistical information. And ultimately - you have the reporting feature, which you can use to plot the progress of your project.

Now that we've seen the types of data that can be used in generating SVTs, our next stop shall be related to the advantages of having implementing SLM in your enterprise.

A contract generally involves at least 2 parties - one offering a service and the other consuming it. The basic idea of a contract is to formalize a business relationship, and may consist of one or more agreements, which define the contract terms.

Agreements in turn may contain language that deal with the specifics of the contract – such as an ETD, cost, options, and so on. These specifics would be the ‘targets.’

To make sure the defined work is getting done properly, there is a need to MEASURE the terms of the agreement. As an example, you can have a contract with a landscaper with an agreement that the work shall be finished in 5 days at the cost of, say, $1000. And if done early, a bonus of $50 would be offered; but if late, then a penalty of $50 would be assessed. This is a simple definition, but it gives enough information on how the work shall be done and the expected rewards/penalties.

In much the same way, IT shops that offer services to their customers have a strong need to be monitored and measured, with feedback coming in via surveys, data analysis, and direct customer comments.

What is the need to measure something like, for example, the amount of time taken to set up a virtual machine for a test environment requested by a QA team?

If you think about it, the advantages of a measurement/monitoring system are tremendous: Scope for improvement, increase in productivity, cost savings, more efficient use of resources, improvement in customer satisfaction – are just a few, as long as the data are being collected accurately and fairly, and are analyzed the right way.

For instance, you can’t fault the IT tech if the VM host itself develops a failed disk – it’s something that can’t be controlled, and thus cannot be counted towards the SLA (the Service Level Agreement).

What an SLM tool should do (terms/definitions/implementations may vary):

1. DEFINE contracts – the overarching buckets that holds everything
2. DEFINE agreements – sets of agreed-upon goals
3. DEFINE and TRACK targets – the actual definition and implementation details of the goals
4. DEFINE and TRACK milestones – how long to wait when an SLA is violated before taking action; or run certain tests at certain points in time during the project
5. DEFINE and EXECUTE actions – when milestones/SLAs are violated
6. DEFINE and TRACK penalties/rewards
7. REPORT on any aspect of the measured data (transforms data into information)

What can be measured? Anything. Absolutely anything, as long as the information about the measured entity can be parsed (mathematically/semantically). As an example, you could measure the amount of time taken to close out a customer issue, and you could also measure the number of times a customer has used the word “terrible.”

Let’s expand on the landscaping example.

CONTRACT – the document you sign to have the crew perform the work

AGREEMENT – that the overall cost will be $1000 and time to finish would be 5 days

TARGETS – cost, time estimate, number of people doing the work, number of bags of mulch, type and quality of top soil

MILESTONES/ACTIONS – at most grant an extra day; notify contractor if delay goes over allotted time

PENALTIES/REWARDS – bonus $50 for finishing before time; penalty $50 for being late

REPORT – give feedback to the contractor when work is done; post review on consumer websites

More later...(especially how things can plug into ITIL processes)

How much sense does it make? Not much. Not much at all. MS is known for its aggressive marketing, product growth, and pushing strongly into areas that have already been cleared for it by others - and very often overrunning the precursors in the process. However, it is not very much known for innovation.

Y, on the other hand, was one of the first true innovators on the Web, bringing a 'directory' approach to search. However, as the Web grew exponentially, people had little time or patience to look through subdirectories and such -- they just wanted the ability to type in something and see something useful come up quickly.

Google satisfied that need splendidly. Its simple, understated interface with just three or four links, and two simple buttons, did it all. Magic, nearly every time. Witness its torrid revenue growth and the merciless streak of profitability, a portion of which comes at the cost of others, mostly Y and MSN (which is, in my opinion, the most anemic of all search engines).

Others somehow stumbled along, while G, with the incredible muscle of its finances and the fantastic brains behind it all, simply left everyone dazed (and tottering).

Little wonder that it cried foul at MS's offer; and even less surprising that it offered a 'helping' hand to Y. But I think secretly G wants MS to get into Y the way a dying man gets trapped in quicksand. Y just announced it would lay off 1000 people worldwide; it has shut down its Photo division, and probably will shutter many others that are simply not contributing to the bottom line. That leaves a WHOLE lot of disgruntled, and in many cases, very talented people just waiting to jump ship.

Enter MS - to hasten the fall, and enter G - to welcome the jumpers.

Y is decaying; I have no doubt about it. Jerry or Terry - same results. Its Panama initiative is not going to get results anytime soon, and worse, MS may cause the most important property of any company - its developers - to quit, thus endangering significantly any future revenues.

Further, while Y has a startup-type outlook, MS is on the other side of the Net divide: Stodgy, self-important, dull, and a penchant for monopolistic tendencies. Therefore, a clash of the cultures is definitely not to be ruled out.

Overall, not a very rosy picture there.

However, as my wife will occasionally point out, not all of my notions are accurate to the last detail all the time. Besides, every now and then I'll come up with a non sequitur or two: Because both MS and Y are competition to G, combined they'll surely kill G.

Yeah, right.

Anyway, the mise-en-scene has been set - let's get the popcorn and watch the fun unfold!

One must remember that desktop virtualization is still new and hasn't really become popular yet, but should take off like a rocket once corporate types figure out it's cheaper, easier, more secure, and more reliable to push out a preformed virtual image to employees' machines than any other solution.

Now that the Server 2008 will have Hyper-V built-in - and with the same OS layer that they sell so much of, the teaming up with Citrix (which purchased XenSource) will further help consolidate their position as being highly committed to the VM platform. VMWare must now fight back with new relationships/partners and technologies that will improve the speed, response, security, and performance of their products in general. Although they have a commanding lead in the market and are seen as the leaders, MS thrives on starting late and catching up then overtaking. So, despite any delay or kludgy/buggy interfaces that one may encounter in MS' first offering, you simply can't write them off. They have the money, the resources, and the doggedness to go after *anything* - however dumb a move it *may* seem to outsiders.

To be sure, the OS is still their main source of revenue; however, they'll take anything they can get in the fresh, still-quite-untested market of VMs. No question it's a new source of revenue (and customers), and it's also one that's bound to grow very fast, and by large amounts. The 'green' message behind VMs helps a lot, plus space and time savings. The Citrix partnership could hold back those companies that want to move from MS to Linux and keep them safely ensconced in the MS fold.

Although analysts seem to be confident about VMWare's current strategy and product direction, they'd do well to keep looking over their shoulders.

All in all, it's a VERY positive announcement from MS, but let's hope they don't come up with another Zune (if they did, then with a little stretching one could call VMWare the Apple of VMs).

You can read more about this sadly misguided person's story here:

http://www.informationweek.com/news/showArticle.jhtml?articleID=205601393

The ultimate irony is that he was kept while other SAs were shown the door.

In previous postings I've mentioned that the biggest threats often come from insiders - disgruntled employees, saboteurs that get employment in the target company so they can perform destructive actions, corrupt workers and so on.

Obviously the idea is not that one should distrust their workers - on the contrary one should trust them completely, but while still taking protective actions, such as routine scans of all admin commands/actions; sweeping the disks of critical servers to check for any obvious problems and so on, and maybe even having a trusted party check important systems for signs of unauthorized or unacceptable activities.

While there are pretty good tools to prevent virus/DoS/hacking attacks, none that I know of protect against such deviously simple yet hard to find attacks. Unless AV software can start incorporating intelligence (singatures) of destructive behavior it won't be simple/possible/easy to stop such people.

Their model is brilliant - simple yet very effective. No question that DELL saw the gem and grabbed it. The only thing is, their website (TNWSC.com) states that they are 'fiercely independent' in the sense that they do not owe allegiance to any one vendor. However, with DELL now buying them out, how does that change things?

From their website's FAQ section:

Check out their FAQ here: http://tnwsc.com/faq's.html (I'm not a fan of apostrophes where they DON'T belong, especially in plurals).

TNSWC do not recommend solutions, yes, and they have a methodology called 'Point of Proof' which DELL is going to market, but still the idea of a previously independent entity flaunting its disinterest now getting bought out by a storage vendor is somehow a bit odd, although I'd think it will make no difference in how TNWSC will continue to work or how DELL will treat its old (and new) customers -- because ultimately credibility (and honesty) is everything. As long as they continue to save their clients tons of money and guide them through the labyrinth of storage acronyms and technologies who cares! I look forward to seeing how DELL exploits this to-be-hot-soon market (that of IT Storage consulting). Companies have invested millions (and billions) of dollars in their complex IT (storage) infrastructures, so if they want to see returns who can blame them! As an analogy I'd say such firms are like the patient advocacy firms - they promise results for your investment; no more no less. See http://en.wikipedia.org/wiki/Patient_advocacy: another hot trend considering healthcare costs and a seeming apathy towards the very people that fund the system - the patients.

****

Another curious thing I read recently related to IBM's release of the semantic search (for email), available on their AlphaWorks site (http://www.almaden.ibm.com/cs/projects/avatar/)

The first thing that'll come to anyone's mind is Google Desktop Search (GDS) - a very powerful and unimaginably quick search tool that I used for a long time before the index became a bit much for the disk (I have a pathetic 12 GB disk). Now, if you had a 100+GB disk with a lot of documents/email etc you'd really want GDS. GDS however does a (I think) strictly string-search approach - no 'intelligence' or 'rule-based search'.

The new tool from the Avatar research team does a lot of similar things -- it mines unstructured information and renders them searchable (albeit in an 'intelligent' fashion -- heuristics, really; so watch out for cognitive biases). So what's new? I know that Stratify (used to be known as Purple Yogi) used to do the same. I think Stratify was funded by In-Q-Tel, apparently the funding arm of the CIA.

The problem statement posed by the researchers/inventors is nothing new - there's a whole lot of information that's just lying there, waiting to be found, associations waiting to be made, text waiting to be indexed. To make the process of digging through the dirt cleaner, quicker, easier, and accessible is an unenviable task. Imagine a corporate website that has individual blogs/mini websites/documents all over the place, containing sensitive, important, and critical material that's probably needed by many others (or they don't know that they need it). An index-and-search tool such as Google's SearchAppliance would be a great thing to have, but only to search for actual strings (again, I think they simply index and search - corrections from the knowledgeable welcome).

With IMB's OmniFind (Omni is overused to the point of being a cliche' now) you could type in, say, 'requirements gathering' and it will search even for something like 'how to create great requirements' or 'the art of successful project management' etc - you get the picture. I'd like to repeat that this is not a new area, but to my knowledge it's also not an area that's been developed very well in the consumer area (including corporate customers). And therefore such initiatives are most welcome as they'll help people do better searches and save a whole lot of time in finding the things they're looking for -- so they can be more productive and efficient.

Not to mention they'll REALLY help trial lawyers when they do e-Discovery (remember, all those rules that you're going to be punching in, creating associations and relationships) could become evidence - not just the results but also the RULES AND THE INTENTION(S) behind the rules as well.

Anyway, I'll give it a try and update this blog sometime next month with my findings on how good it is.

http://www.informationweek.com/blog/main/archives/2007/12/we_need_to_talk.html

John Soat talks about how end users take untold liberties with IT policies and probably take them as suggestions rather than mandatory rules. It's quite complex as to why this happens -- it's quite clear that they probably know what they're doing is wrong, but just not *so* wrong that they shouldn't do it.

See, the issue is that many of these areas are left in the gray part of 'can do or must not do' policies. And worse, I'm willing to bet that 99% of employees have NO IDEA what constitutes proper 'secure' behavior and what constitutes a violation of company policy and thus their employment contract.

Along with continual education, the only other way to make sure that corporate data doesn't leave the network is by using software to track the packets and ensuring they are not sensitive. To do that you'd have to get one of those 'anti leak' DLP software modules (like what Vontu/Symantec does) and establish clear demarcations between acceptable and non-acceptable information leaving the network.

Any practical ideas from readers?

A slew of announcements this year must have gladdened the hearts of those that store away every last bit of data, those that don't/won't delete a single temp file.

With Microsoft's SkyDrive, Google's "GDrive", Amazon's S3, EMC's purchase of Mozy and many others now making a push for online backup/storage and promising ease of retrieval and ubiquitous availability of data (all you need is a browser), suddenly your D: drive doesn't look so big/small anymore - depending upon how you see it.

Some of these sites have automatic backup features while others are simply 'dumb' drives that you upload files to - but the overall principle remains the same: Stop clogging your disk space with files that you can store online instead. For some others it might be a dash of paranoia in storing the file in multiple locations, and for some it might be a case of hey-there's-nothing-to-lose-and-it's-cheap!

In any case, what all of this boils down to is this:
Ordinary computer users are now becoming storage-savvy. The falling prices of storage devices, the dramatically reduced form factor, availability of advanced technology combined with an intuitive interface - all have made consumers demanding and resourceful.

As more and more aspects of our lives go digital, what used to be a "WOw! 250 GB - what am I going to do with all that space!?" is now "I have a NAS that can expand up to 2 TB."

I am thinking that commodity storage will both help and hurt storage vendors: help them by driving up demand in the online storage world, and hurt them by turning off the nascent consumer market, one that probably is much more profitable and much less exacting. However, who's to say both can't co-survive? They very well might - and that's what the future will tell us.

So, turning now to the online storage guys -- how will they manage this demand? I'm hoping the solution won't be to throw more disks (JBoDs) at the problem but will involve meaningful and relevant storage management, starting with some solid prediction models of resource usage.
Without that, where would one even start?

Next would be evaluation of various technologies - SAN/iSCSI/SAS/NAS - which? Maybe all of them - they all have their advantages and drawbacks. But for the sheer scale of they're trying to achieve (couple that with disaster recovery - with the DR center most probably located several hundred miles away) and you realize the enormity of the issue. Again - the fact that disks are cheap doesn't help but only adds to the problem.

So, a sane person would sit down with reputable consultants, spend decent money on experts/consultants coming up with an architecture that satisfies the criteria the CTO (and/or Legal) is asking for, then develop a quick prototoype to see how things fit. In this stage evaluating various vendors would be a good idea for benchmarking, pricing, scalability, and affordability.

A totally worthwhile investment would be good SRM software.

However, none of these will really bring any results unless there is a SOLID PROCESS to hold the entire thing together. Here's where ITIL and products that help implement ITIL-related methodologies will assist tremendously.

Check out various systems-, network-, and storage-management software to see which one would help you most. But beyond that, you really need to tie them with the help of a CMDB, and that's where a BSM (Business Service Management)-type solution can prove invaluable. The idea is to link the failure of a disk to the exact business unit that's affected (say Image Services) and then calculate any losses/SLA violations, and also figure out any compliance issues.
Such a system is surely going to be expensive at first sight, but the benefits will soon far outweigh any cost concerns.

Which would you rather have: A $3M outage or a $200,000 software that'll save you that $3M and more and more on a perpetual basis? Of course, what I mention is hypothetical, but such a situation is not uncommon at all! In fact, the costs of fixing a data breach or a downed network keep going up because the speed of business is getting only higher/faster and each 'unreachable' or '404' or '500' probably means several customers lost. Maybe several hundreds/thousands including existing ones that surely will go for something more reliable.

...and I haven't even gotten into security yet...

http://sunbeltblog.blogspot.com/2007/11/breaking-massive-amounts-of-malware.html

I thought it was a minimal but visually arresting article - enough information to make sure you don't stumble into the dark areas of the web - or at least know what to look for.

By seeding all sorts of sites (blogs/trackbacks/comments) with their infernal site links they try to fool search engines into listing their URL at the top, or at least at the middle of the search results. Unwary users will no doubt not bother to CHECK the URL before clicking it, and what happens next should not be surprising: a whole lot of popups for installing malware/rootkits/password stealers, and of course, the maddening ads.

Here's my suggestion:
When you search for anything, first make sure you check the URL to see if it's a nonsensical mix of meaningless words. If yes then stay away.
You could also try searching your favorite sites first (such as GPSPassion/Poi-Factory for GPS stuff; ExpertsExchange for technical questions; dpreview for camera questions etc). You get the idea.

http://www.guardian.co.uk/feedarticle?id=7092429

I noticed that fellow blogger Jeff Bohren also mentions it.

Seems like an ingenious way to make sure that the next incoming plane is exactly that - a plane and not a missile (user authentication).

The idea seems pretty straightforward. Israel issues these devices - called SCS/Code Positive - to every airline that wants to do business with it; the pilots use them to authenticate themselves upon approach. How exactly? Well, details are not available, but it looks like a credit card-size, calculatorish-looking, geeky little tool. I guess they'd enter a code/speak into it/something else.

I guess somehow at some point the pilots are given a number that they need to enter into the SCS when they're about to enter Israeli air.

The actual working is probably known only to the company that made it (Elbits), the pilots, and maybe a few other folks.

It addresses 2 types of situations:

a. Hijackers have disabled/killed the pilot and have taken over the plane

b. Hijackers have a gun to the pilot's head

What about the third - that the pilot himself/herself is a terrorist (belonging to a sleeper cell)? What then?

So, if the code does not match the pilot will get several warnings, followed by the IDF planes doing a 'what's up' up close, and then followed by (upon ignoring their orders) a hello from the IDF's air-air missiles.

Unless we know for sure how this little gadget works, I think it's a lot of fun to speculate, as long as we're not on either side of the device!