Skip to content

All-in-a-box vs. Big 4

The White Boxes

I talk with the press frequently. They thankfully whack down my ramblings into concise quotes. For those who prefer to see more, I try to dump publish slightly polished up conversations I have with press into Press Pass.

Barbara Darrow of TechTarget recently asked me what effect unified, single-stack servers could have on the existing IT Management landscape. If these offerings, like Cisco’s UCS are really “everything you need in a box,” what place do traditional, 3rd party systems management vendors have? Here’s my response:

When it comes to unified devices, I don’t see the dynamics changing dramatically. At the moment, while things like monitoring and event management are required to run IT, they’re not big cash cows for large systems management companies compared to higher level functions like service desks, the process around service management, reporting, and keeping up with virtualization and self-service portals (which is what a lot of what “private cloud” ends up being).

One of the major problems with these bundled systems management offerings is that they tend to only work well with the hardware or software stacks they’re bundled with. You still need to monitor all that other stuff you have: like your Microsoft software, your routers, SAP installs, Unix boxes, etc. I typically see these embedded systems management offerings used as the best way to drill down and diagnose a problem on that system [the end-to-end stuff demo’ed at SpringOne 2010 from VMWare/SpringSource is an excellent example], but the staff tends to use a general purpose monitoring tool to keep track of everything in IT on a day-to-day basis. There’s also a tremendous amount of concreted-in legacy that requires monitoring, despite what fancy new “cloud in a box” machines might offer. Of course, if all you have as a UCS or the like, you may be set.

Raw monitoring hasn’t been an incredibly hot area for sometime, so there’s not a tremendous amount to loose. The more important turf is the software around helping IT manage change requests, keep track of assets, and otherwise assist in the processes and workflows IT goes through day-to-day…and then reporting on all of that for upper-level management who need all the muck below them summarized in a dashboard, or report, to help make decisions about what to keep doing, what to starve or kill, and new IT services to start.

Others like IBM Tivoli are smartly moving into other industries (like monitoring power grids, city infrastructure, etc.) where new networks are emerging and require monitoring afresh. If everything is IP-addressable and on a network, you can tweak your data center monitoring tools and apply them to, say, large power grids or even “nodes” in waste-water management. This last point is pretty unique to IBM.

Cisco has partnered with BMC around UCS, which is evidence of BMC keeping “in the game” despite Cisco trying to capture more of the data center dollars.

Buyers tend to maintain a healthy skepticism about buying everything from one vendor: the other side of “one throat to choke” is “one big check to write.” Instead, buyers tend to still like the ability to pit vendors against each other to drive down cost and get better service.

But, to your point, all of the vendors are going crazy for this one check to write approach: Oracle would love to see people buy their entire stack, as would IBM, and Dell’s VIS are all going towards this idea.

Disclosure: IBM and Dell are clients, as is Microsoft.

Categories: Enterprise Software, Press Pass, Systems Management.

Tags: , , , , , ,

Comment Feed

5 Responses

  1. [disclosure – I work for Cisco]


    I’d agree with you that concern about lock-in or future loss-of-capabilities is often the first reaction people see when bundled systems are released into the market. But in the case of Cisco UCS, I believe that you’ve mischaracterized where the integration starts and stops. While Cisco UCS does have a single point of management for the Chassis, Server, Service Profile and Access Network, the management framework doesn’t lock it into any single management system. Everything that you can manage on UCS from the UCS Manager (GUI), you can also access via the UCS API (XML). This allows 3rd-party systems – BMC, CA, Tivoli, HP Ops Mgr, Microsoft System Center, newScale, Dynamic Ops, etc. – to manage UCS as part of broader IT operations.

    When we designed UCS, we believed that convergence of Compute and Network would happen quickly due to Virtualization and a movement towards Cloud Computing models. We believed that we had the technology and innovation to bring those elements together in unique ways. What we didn’t assume was that customers would immediately replace their existing operational systems and process, since they have been in place for years. Hence the reason we designed UCS the way we did. Simplify the infrastructure, simplify the management, but don’t dictate how it will be managed. Based on customer and partner adoption rates, we’re finding this approach to be quite successful in the market.

    More details can be found here:

    We’d be happy to discuss the details of UCS management in more depth if you are interested.

  2. Brian: thanks for all the details 😉

  3. Cote, lets re-state the Dell VIS perspective. We will ship on an all Dell hardware stack as the management system. However, we will definitely support multiple hardware vendor servers, networking and storage as managed endpoints. In R1 we will be limiting the number of these, but we will definitely be supporting Dell, HP and IBM servers, as well as numerous Dell and non-Dell storage and networking as managed endpoints.

    We are committed to open since we don't have all the points in the stack to be able to lock people in.

  4. Mark: thanks for fleshing it out!