The Organizational Impact of Converged Infrastructure

One of the ways that companies are looking to become more agile is by breaking down the silos between groups.  This change can not happen solely through the adoption of new technologies (including management tools), it requires that IT staffs look to cross-training and internal process changes to become more efficient.  Virtualization is a catalyst for this change as server, network, storage and application owners are all dealing with the impact of abstraction on how they do their jobs.

At the opening of an SAP Center of Excellence in Santa Clara, CA (includes a shared lab with the VCE Coalition), Chinh Van of Calloway Golf discussed the impact of convergence on his organization.  There is a blurring of network, system and storage administrator roles.  Chinh mentions (see the clip below of the full interview here) that a critical piece to making this happen is a single tool that can manage across these domains.  The promise of a “single pane of glass” that can manage the environment is attractive, but there is still a need for the knowledge of domain expertise that is currently held by various groups.  Network administrators understand how to provide bandwidth and scale an environment, storage administrators make sure that there is proper data availability and data integrity; even with convergence and virtualization, these attributes are needed.  Some fear that the goal of convergence is to eliminate jobs, but the reality is that if companies do not change the way that they do things, they will not be able to keep up with the huge growth of data.

One of the best ways for groups in an organization to start working together is to be put in the same group where activities and goals can be closely aligned.  Cisco and EMC took a similar approach with server, storage and network products when they put together the Vblock architecture.  What started out as a reference architecture (which VARs would build) over a year ago is now shipped from the factory as a single SKU.  David Floyer of Wikibon posted a financial analysis of the Business Value of Integrated Stacks.  The immediate advantage of a fully integrated solution of servers, network and storage is that implementations are simpler; racking, cabling, and installing software is all done at the factory.  The larger savings is that operationally, the system is treated as a whole rather than managing the pieces.  Specifically, Vblocks are managed using UIM (EMC Ionix’s Unified Infrastructure Manager), what EMC’s Chad Sakac calls “orchestration purpose built for Vblock”.  Managing the servers, storage and network is not as simple as vendors say (management has long been listed as one of the top challenges), the pre-configuration and simplification of the environment is easier to manage than legacy environments and is a move in the right direction.

Before organizations can fully exploit converged infrastructure they must consider the organizational issues convergence brings. Who is ultimately the leader responsible for managing the infrastructure – is it the server, network or storage team? Are these individuals on the same team? How do you properly incent them? What is the reporting structure? How is escalation handled and what types of decisions can be made without escalation?  Wikibon has discussed this extensively and believes that while the network teams may increasingly gain power  within organizations – especially as Ethernet adoption grows – the bottom line for storage is that storage professionals must still ensure that data is not lost. Networking and storage teams have different mindsets and before organizations mash them (or their infrastructure) together they need to be on the same page – or at least reading from the same book.


, , ,

  • Pingback: Tweets that mention The Organizational Impact of Converged Infrastructure « Wikibon Blog --

  • Anonymous


    A great topic for a post and one that does not get enough attention. One of the things I always bring up in my exec briefings is that if you are going to re-architect the data center for things like virtualization, convergence, cloud, you need to re-achitect the org structure to match.

    So, a couple thoughts on this beyond the ones you bring up. First, the technology can play a role in making this helpful or painful. From our perspective, we have always tried to be non-disruptive on the operational end–having converged infrastructure, but leaving the mgmt tools and operational processes unchanged, so a SAN admin can approach the N5K like a fabric switch and network folks can approach it like a LAN switch–on top of this we layer role based access controls. But this only gets you so far–there still needs to be governance about things like maintenance windows, etc.

    The one other area to address is making sure folks understand their role in the new converged architecture. With unified fabric, folks often assume someone’s job is going away, which is actually not usually the case. With converged infrastructure, the transport is getting unified, the services themselves are not. The best example of this is IP telephony–the dedicated voice infrastructure went away, the role of the voice team and the need for their expertise did not. From a leadership perspective, this might be self-evident, but its still important to have an explicit conversation with the affected teams.


    Omar Sultan

  • stu

    Thanks Omar – excellent points. It’s a nuanced message – things are going to change, roles will be shaken up, but overall it can be a good thing as long as at the end of the day people can get more done without risking the integrity of the solution.

  • Stu:

    Indeed, organizational structure has lots of intertia, so it often adapts much slower than technology can. Crisis really reduces that inertia, business crises in particular. What business crises do you for see that would cause a CTO/CIO to go through the painful turbulence of changing the organization?

  • stu

    Hi Brook – take your pick of crisis: data explosion, companies spending 70% of their time “keeping the lights on”, consumerization of IT (or the “Google effect” if you prefer). Just as we say a huge shift in IT from the mainframe to client/server, we are in the midst of another transition . The change does not happen overnight, but there are definite competitive pressures to understand where you are today and what changes need to happen.

  • I see. Would you agree that in this context “server virtualization” is a proxy for a centrally managed and controlled datacenter infrastructure whereby IT skills can more simply coordinate their actions?

    The point is not that a skill is eliminated, but that skills need better coordination platform to be efficient. Like the mainframe management model (which is often denigrated by the ignorant), we need an x86 management platform that coordinates the actions of work load management, network configuration and storage administration, etc. We have it for the MF and UNIX, we are missing it for x86.

  • I talked about this a long time ago called “Collapse of the silo” and some believed in it, some were disinterested, and some downright hostile.

    The most hostile people I found were middle management in charge of a team of single-skill specialists, and the single-skill specialists who are comfortable thankyouverymuch.

    You can see my prezi here Currently reading Collapse of the Silo

  • Alex

    thanks steve for ur participation.


  • Pingback: Accelerating the Industry Towards Converged Networking « Wikibon Blog()

  • Pingback: HP Touts “Open” Approach to Converged Infrastructure Deployments | SiliconANGLE()

  • Pingback: What does the future hold for IT skill sets? | By The Bell()