Share:|

iceberg.jpg

Hardware monitoring is traditionally such a tough task that some IT departments avoid it altogether and prefer sending staff walking around in the server room to check the color of the LEDs in the server racks! That’s the prehistory of server management...

 

Traditionally, in less-prehistoric companies, big IT environments rely on the hardware monitoring tools provided by the server manufacturers. The immediate advantage is that such solutions come free of charge. Apparently so. The challenge is that the more different platforms they have, the more complex the monitoring of these servers gets.

 

WHY FREE VENDOR-SPECIFIC SOLUTIONS AREN’T SO FREE

Hardware vendors (well, most of them) provide some sort of hardware monitoring solutions in the form of agents and consoles. Windows administrators all know HP Insight Management. Sun administrators all know Sun Management Center (SMC). Some people may also mention IBM Director, but it’s far less common.

 

Those solutions are available for free. The issue is that each vendor-specific monitoring solution is specific to one vendor. Sometimes it’s even worse: it’s specific to one server series from one vendor!

 

HP Insight Management, for example, was specifically designed to monitor HP ProLiant servers (actually, it was Compaq at the time, but let’s forget about prehistory). But HP Insight Management cannot monitor IBM xSeries servers. Neither can it monitor IBM AIX systems, nor Sun Solaris servers, nor Dell PowerEdge. It is even bad at monitoring HP-UX systems, as it only integrates with HP-UX’s STM (Support Tools Manager) which fails to gather critical information from the out-of-band management card of such systems (status of power supplies, temperature sensors and fans, notably). And while HP claim they can monitor Dell or IBM servers with HP Insight Management, it is only as a basic SNMP MIB integration. You will then still have to define which SNMP variable is important and the meaning of each possible values of such variables. Painful.

 

The case is easier with IBM Director and Sun Management Center as they don’t even pretend to be able to monitor something else than their own servers.

 

It gets worse. Not all manufacturers have something similar to HP Insight Management. What about IBM AIX? What about Fujitsu-Siemens, NEC and Dell? For such systems, administrators have to write their own scripts to check hardware failures on their servers.

 

All companies have servers from different vendors: HP, IBM, Dell, Sun, NEC, Fujitsu. These companies need to integrate all of the corresponding vendor-specific hardware monitoring solutions (either provided by the manufacturer, or written on their own) into their IT management systems: monitoring consoles, notification systems, incident and problem management, asset management, etc.

 

So, if we take a company with HP ProLiant, HP-UX, IBM xSeries and IBM pSeries (AIX) servers, and this company already put in place an asset management system, an incident management framework and a notification system, the integration diagram would look like that:

free_hw.jpgUnfortunately, when considering about going with another hardware vendor, like Dell or Sun, it is important to have in mind the price of integrating the monitoring of the hardware of these new servers into the asset management, incident management and notification systems. This major fact adds to the complexity and has a cost.

complicated.jpg

So, this “architecture” gets more and more complex for each new vendor or type of server that has to be integrated.

And this gets even worse if a company wants to add a new event management system. Because all of the vendor-specific monitoring products need to be integrated with the new component. This “free” solution is more and more looking like it has some not-so-well hidden costs!

cost.jpgActually, some companies do realize that and avoid purchasing new cheaper servers from another vendor simply because they won’t be able to integrate these servers in their management tools at a reasonable cost!

 

That’s exactly what some server manufacturers want: tie in the customer to their specific platform so they can’t go get cheaper prices elsewhere. That’s what we call a lock-in. And this also explains why such server manufacturers will never provide a solution able to easily monitor servers from their competitors.

 

BMC platforms combined with the Sentry Software’s products form the only monitoring solution that is truly vendor-agnostic. Servers from Dell, HP, IBM, Sun, Fujitsu and NEC running Windows, UNIX or Linux are equally monitored in the BMC TrueSight Operations Management (formerly known as BPPM) environment.

value.jpgIf the IT department wants to buy a new type of servers from a new vendor, they can be assured that it will be covered with BMC TrueSight Operations Management and, in turn, will be fully integrated into the bigger picture with the incident management, the asset management, the event management, etc.

complexity.jpgAs a proof from the market, the largest IT outsourcers, whose job is to manage the heterogeneous IT environment of their customers, all rely on BMC Software and Sentry Software solutions to monitor the diverse hardware platforms of their customers.

 

SO, HERE IS HOW THE THINGS COMPARE:

Patchwork of vendor-specific monitoring solutionsBMC + Sentry
warning_16.jpgComes free of chargeli-green-check.pngLicense fee
warning_16.jpgGaps in the monitoring coverageli-green-check.pngCovers 100% of the IT infrastructure, including storage!
warning_16.jpgHidden cost: Painful integrationli-green-check.pngEasy integration with the management tools
warning_16.jpgHidden cost: Avoiding new (cheaper) vendorsli-green-check.pngEasy integration of new vendors
warning_16.jpgHidden cost: May require custom developmentli-green-check.pngOne single standard software solution
warning_16.jpgHidden and therefore hard to evaluate costs: budget and cost planning nightmareli-green-check.pngCosts reduced to almost just the license fee

As always, beware of the hidden costs behind so-called “free” solutions!