Managing and protecting all enterprise data


Manage Learn to apply best practices and optimize your operations.

Best Practices: Why you can't go it alone

Most IT departments are split into islands of expertise, like storage, servers and networks. But the time is coming when those disparate groups will have to learn to work together. Network consolidation is just one of the technologies forcing storage pros to build stronger interdisciplinary IT teams.

Network consolidation is just one of the technologies forcing storage pros to build stronger IT teams.

BACK IN THE early days of computing, there were do-it-all IT teams to manage your system, network and storage. Then technologies became more complex, and it was more efficient to create silos for network, storage and systems administration. The storage admin role that we know today was born out of the need to manage storage as an independent entity--online and offline, including backup and recovery. The same could be argued for network administrators who were tasked with managing these new environments, making networks secure while connecting them to the powerful and potentially dangerous Internet. Systems admins were left to manage the rest. Recent advances in the technology landscape, however, may signal that history is repeating itself.

Recent acquisitions by Cisco Systems illustrate this point. Remember when Cisco was known as a leading network solutions provider? Now the company is an industry front-runner in Fibre Channel (FC) and IP SAN in large part due to its acquisition of Andiamo Systems. By "platform sharing" between its IP and SAN products, Cisco delivers a multiprotocol uber-device that has Layer 2 and Layer 3 capabilities for TCP/IP and FC stacks. This allows administrators to have a common platform for IP and FC SANs. Cisco's recent acquisition of Nuova is another indicator that it's serious about creating a single, unified access mechanism (akin to InfiniBand) for transparent sharing, provisioning and administration of everything from server access all the way to the application I/O stack. Other vendors are traveling down the same path. For example, Brocade recently acquired Foundry Networks. A few years ago, this acquisition might have left everyone scratching their heads. Today, it makes more sense as we see vendors consolidating the I/O and access stack.

From an IT-preparedness standpoint, it's not too early to start thinking about how IT silos or teams at your company cooperate and collaborate as new technologies merge.

New attitudes for new technologies
In a flat organizational chart, all teams should be considered equally important. This allows for greater cooperation and collaboration. In the same way, new technologies shouldn't be treated as a specific team's responsibility. Is your virtualization project the property of the storage team or the server team? People still need IT managers. But that manager needs to ensure that they're perceived as an equal (not superior) member of the team; more importantly, they need to guarantee that each team has all of the information it needs regarding IT operations and expectations. It's called transparency and it's important to your overall success.

Consider creating a cross-functional architecture team. It helps system architects stay in touch with the operational reality of the solutions they design. It can also create an IT team that considers issues related to all of the functional teams impacted by their work. A case in point is boot from SAN with its quick deployment of multiple images of the same build or platform, quick patch rollback mechanisms, point-in-time copies and replication of the OS. As challenging as it may sound, only complete collaboration between the storage and systems folks can make this project a success.

Perhaps server virtualization or blade server deployment is the single largest catalyst for changing the way we view IT roles and responsibilities. Virtualization is somewhat analogous to how boot from SAN was treated when it first became feasible for OS file systems. These systems were traditionally stored on local disk on the SAN, a spot previously reserved for application data. Suddenly, a new world of possibilities emerged. Virtualization is changing things in a similar way. For example, a favorite location for storing guest images is on NAS. This means provisioning, supporting and troubleshooting for virtual systems on NAS already involves the systems, storage and network teams. However, most systems currently have dedicated access paths for FC and IP, respectively. You could therefore argue that each team has its own access paths to manage and maintain the system. For example, when the network team configures a network interface, the storage paths (via FC) aren't affected. Assuming industry predictions prove true, there will soon be a single port on the back of the server that will be used for both types of access. Now you have the network team configuring a network interface that may have the potential to affect storage traffic. That could be tricky.

Add storage virtualization to this scenario and things begin to look downright dangerous. Today, storage virtualization enjoys the benefit of a dedicated FC or IP network (for NAS virtualization). Tomorrow, with a unified network, that may no longer be true. The teams managing tomorrow's network will also need to be connected in new ways.


Change control and cross-functional rotation
To get a better understanding of how current IT teams function, take a look at your provisioning process. If possible, run a report to see how long each team takes to perform its provisioning task. Also check to see whether the hand-off process is streamlined or if the tickets jump back and forth between teams for clarification and other miscommunications. This will give you insight into how each team functions relative to others and allows you to see any organizational, functional or operational challenges.

The same policy applies to change management and control. Check to see if a particular change activity or task can be performed without stalling repeatedly. If there are significant holdups, determine if they're being caused by a particular team or process. IT teams should have incentives to educate other teams about their processes, and even a procedure for allowing one team to critique another. Some of your competitors are already doing this in an attempt to provide better IT services to their business.

When IP SANs came along, storage admins faced a challenge in acquiring TCP/IP know-how. Likewise, network admins were faced with the question of how to increase their understanding of storage and apply it to their networking systems. IP SANs--and to some extent NAS--have provided immense training opportunities for storage and network admins to be able to stretch outside of their comfort zones and learn different technologies.

But new technology training doesn't have to be costly. For starters, it doesn't mean that you need to send your teams en masse to vendor training sessions. If your firm has an internal training group, you could collaborate with them to create your own course materials and labs. Team members can also contribute toward developing specialized course material. For example, systems admins could create a course on basic Unix administration for the network and storage folks.

As this concept matures, you can even involve members from different teams in the systems, storage and network on-call rotation. This allows everyone to get a taste of the challenges faced by their peers. For example, systems admins will learn to appreciate the extremes of supporting a storage and network environment. If you try this, I'd suggest a buddy system in which a more experienced person is paired with someone just learning a new network technology.

By now, you might be rolling your eyes about all of this feel-good team building filled with buzzwords and buddy systems. But there's a reason collaborative project tools are purchased so frequently. There's also a lot of good data being supplied by IT teams who have tried a new approach, and there's no doubt your overall data center strategy will require more cooperation and collaboration between people who might never do business together. Consider your CIO, the person paying your monthly electric bills or even your property manager. Each one has information the other one needs, and for a truly next-generation system to work, you need progressive IT leadership where everyone works smarter, if not harder.

Article 18 of 18

Dig Deeper on Data storage strategy

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.

Get More Storage

Access to all of our back issues View All