|
Dependency Mapping - snake oil or success? |
|
Written by itsmbuzz
|
|
Oct 27, 2009 at 05:23 PM |
|
With the recent acquisition of Tideway by BMC software, the last of the independent application dependency mapping tools is off the market. nLayers was bought by EMC, Collation by IBM etc etc. But do these technologies really provide value, or were they more of a 1.0 technology that everyone (especially the analyst community) expected these vendors to have?
The concept of dependency mapping is extremely compelling. Wouldn't it be great if you could have an accurate map of how technology supports a business service, updated in real time? Wouldn't that be an amazing resource to do change management, business impact management, fix production problems and so on. Wouldn't it also be great if these vendors could make their CMDB work by magic - to really have that silver bullet that they have been marketing?
However, the reality is vastly different. In real live environments - with really complex, customized applications that span many tiers, have many levels of technology, are highly virtualized and span over multiple data centers, these tools struggle to provide an useful map of how this mass of infrastructure actually supports an application. Sure, they are great at finding all the servers, but a simple nmap scan can do that. They can even find some basic components, like databases and web servers, but again, nothing rocket science about that either. Really understanding the application is beyond their reach.
A couple of examples.
I was associated with a BSM project for a large manufacturing company. This company had invested millions in a CMDB project and was trying to populate this CMDB for impact management and change management. Given the size of company, the chaotic environment, that spanned over multiple service providers, a manual approach to the CMDB was necessary. The discovery tool did find a bunch of 'stuff' - a very bottom up approach where a dazzling number of hosts were found. This would have been great if asset inventory was their goal. They also found some base components. But the applications were custom, so they were not discovered. Instead, the tool generally identified that certain servers talked to each other. Even making the assumption that the servers that were constantly communicating with each other must be part of an 'application', it was still very difficult to work out what the application actually was (especially in a highly leveraged environment), and how these components contributed to the application. The result was worse than nothing at all, because the data was very incomplete, and most of what was seen was shared services that were hooked into all applications. This company started again on the CMDB, and we moved forward with BSM without it and completed our project.
In another project, an online company was trying to identify their core applications and what made them up. Their applications were all custom developed. This company used a leading discovery tool, but found that creating, maintaining and updating the definitions that were used to discovery these applications was so much work it was easier (and much cheaper) to hire a few interns during summer holidays to manually go round entering data into their system.
Finally, today, I met with a large organization that derived a huge part of the their revenue from a set of key applications. Understanding the dependencies was seen as an important risk mitigation strategy. Another leading discovery tool was selected (a different one from example 1, or example 2). Months later, the project was abandoned after a series of half completed infrastructure maps were created.
I could continue with these examples, but it seems clear from real world examples, that unless you have very simple standard applications like MS exchange, at best these tools take a huge amount of work to get going? Is it worth it? Obviously some think that they are.
Personally, I would rather wait until the second generation of these tools becomes available that can deliver real value - and focuses on identifying the business applications, not blindly discovering basic infrastructure. |
|
|
History is written by those with the largest marketing budget |
|
Written by itsmbuzz
|
|
Oct 27, 2009 at 08:39 AM |
|
Does a claim of having invented something really provide you with that much credibility in the market, especially when its patently obvious that the invention was done elsewhere?
Having looked at other industries, its rather common for the 'inventor' to be replaced by newer companies who add efficiencies to the idea and do it rather a lot better. One could argue that the first airline to pioneer travel to the masses was Pan American, yet they closed their doors in the early 90's, out maneuvered and surpassed by more innovative competitors. Or take the car, 'invented' by Mercedes Benz, but today are a niche player in todays car industry (albeit a well regarded one) - but few buy a Mercedes Benz because they made the first car. What is probably more interesting is how Ford revolutionized the car industry with the production line.
looking back to IT Service Management, there have been many claims over the years to have 'invented' BSM. BMC software is the most guilty here, which is rather humorous since they were a comparatively late player to that market. Managed Objects (before the Novell acquisition) were another, yet again, IDC had written about the concept long before MO started doing so. Is reducing your credibility with these claims worth the added value? I guess BMC and MO thought that it was. I wonder how many actually bought because of this.
What might have been more interesting would have been if BMC made the claim that they had revolutionized the BSM market space. In reality they probably did. They took an established market segment (although small at the time) and twisted it to put the Configuration Management Database at the center, and closely align it with the service desk; conveniently the two strengths of BMC. This focus on CMDB has been to the detriment of the BSM concept, but certainly vital for BMC's sales success over the last few years. Certainly now the 'what is BSM' discussion is confusing and fraught with religious style debates about the merits of the CMDB, discovery and so on.
But I wonder if they would have won more Kudos if they had come out and correctly said that they took the CMDB to BSM and how BSM was incomplete without? (even if in reality making the CMDB the centerpiece of BSM is akin to using a sledgehammer to hang a picture)
Of course none of this really matters and at the end of the day its more of a philosophical debate, but interesting nevertheless. |
|
|
Who's Hot in Business Service Management (Part 1) |
|
Written by itsmbuzz
|
|
Jul 11, 2009 at 07:17 AM |
|
Forrester recently published a report on Business Service Management, which was an update on their previous BSM wave, as significant changes had occurred in the BSM vendor landscape.
Like the wave, while the report was very comprehensive in that it covered a lot of solutions, it didn't really give much guidance as to which BSM solutions are suited to what purpose, and which are really doing something else, but are branded as BSM. Indeed, they listed every solution they reviewed in the 'leader' quadrant during the wave, which reduces their credibility.
The criteria they also used for this report, was somewhat lacking as well. Forrester take the view that a Configuration Management Database (CMDB) is an essential part of a BSM solution, and infer that it is even a prerequisite. I think including the CMDB as a prerequisite misses the whole point. Unless the BSM solution is being driven from the business and the top down, its going to be nothing more than Event Management on steroids. The problem with relying on the CMDB is that invariably the solution moves from being a top down view of what the business needs, to a bottom up grouping of infrastructure into business containers.
What would have been a more appropriate pre-requisite is end user experience monitoring. Too many of the BSM solutions look at things through the point of view of the infrastructure and its vital to consider how the applications/business services are supporting their users.
However, given that Forrester used such rose tinted glasses to rate the vendors, here are a few thoughts of my own:
The Big Four:
The big four solutions all lean heavily on the CMDB, which makes implementations extremely long, and replacement very difficult. Take into account that if you go with one of these solutions, it will take significant time to implement (unless you already have much of it) and be difficult to replace later on. Thats not to say that you can't get value from these solutions, but be aware of the downsides. The upside is that these vendors have a tool for pretty much everything.
HP Software - Probably the strongest of the Big Four in the BSM space. HP have a good foundation with the combination of the Mercury and OpenView tools. Their real end user experience solution (RUM) is a weak link in an otherwise good portfolio of solutions. Like the rest of the big four they leverage the CMDB heavily and require many products all to be integrated together. A good choice if the organization is predominantly HP, but be mindful of vendor lock in and long implementation times.
BMC Software - BMC have exceptional marketing and thought leadership in the area of BSM, and have done some extremely good acquisitions over the last couple of years. Their current direction would appear to be more around data center automation. However, their approach is heavily leveraged on their Atrium CMDB, and the solution which actually provides much of the BSM functionality (Service Impact Manager) is the weak link in their portfolio. BMC are a good choice for those companies who have already invested in Remedy and Atrium and already own a significant portion of the BMC 'stack' (which has been cynically branded as BSM). BMC is not a good choice for those who are not fully bought into the Remedy/Atrium message as getting a CMDB stood up is a multi- year project for most, and the two significant holes (SIM, and a poor End User Experience offering) make other vendors look more attractive.
IBM Tivoli - Not at the same level as BMC or HP when it comes to business service management. Their event Management solution is particularly strong with the NetCool acquisition, but this hasn't really translated all that well into Business Service Management. Likely only a consideration when Global Services are involved.
CA - CA do not really market to BSM, and they are definitely the weakest of the big four in this area. Their focus seems to be digesting the various acquisitions they have done over the last few years. Their application performance solution (Wily) is completely separated from the rest of their business which makes providing an integrated solution far more difficult.
More specialized vendors coming in part 2. |
|
|
|
BSM
