8.2 Qualifying and Quantifying IT Deployment

Faced with the requirement to keep storage up and available all the time, IT professionals need to continually qualify and quantify new deployments. Earlier in the book, we covered the trend towards defensive storage purchases, particularly the focus on disaster recovery solutions. In this section, we cover offensive approaches to evaluating storage projects and focus on measurement techniques that take a more strategic view of the overall enterprise infrastructure.

8.2.1 Minimizing Improvement Costs

To measure and minimize improvement costs, the IP storage model can be used to categorize the role of the IP core across storage deployments. Overlaying the intelligence areas of host, fabric, and subsystem with the access, distribution, and core layers allows IT planners to map solutions that can best take advantage of IP storage networking in the larger enterprise context. As shown in Figure 8-4, a variety of current and planned storage platforms can be mapped to the IP storage model. For NAS, servers directly access NAS filers across an IP and Ethernet core . For storage consolidation applications, iSCSI servers might connect to an IP-to-Fibre Channel router to consolidate on enterprise storage. In the remote replication case, Fibre Channel disk systems are linked to each other across an IP and Ethernet network.

Figure 8-4. Overlaying host, fabric, and subsystem intelligence with the IP storage model.

Looking at the broader enterprise storage landscape, we see the emergence of product categories across the primary storage platforms of SANs, NAS, and object-oriented storage. As shown in Figure 8-5, each storage platform has a variety of products that make up the complete solution. These products fit across access, distribution, and core layers. Let's take the example of a conventional Fibre Channel SAN. In a typical scenario, a customer might have Fibre Channel servers and disk arrays. These products would fit into the access layer. A SAN appliance, such as an intelligent storage node for virtualization, would fit in the distribution layer. Fibre Channel switches would fit into the core. Note, however, that the core of this product suite, being Fibre Channel “centric, cannot be consolidated with other product cores from NAS-or object-based storage, as those cores are built with IP and Ethernet networks.

Figure 8-5. Minimizing improvement costs through product consolidation.

This kind of infrastructure outline can help when planning platform consolidation (through product consolidation) and also in balancing storage intelligence across various layers. Let's dig deeper into the example of remote replication for a SAN. A current solution might include two Fibre Channel disk arrays with subsystem intelligence for mirroring and replication. These intelligent arrays include all of the required software to make use of transparent distribution and core layers. A new solution might migrate some of the remote replication functionality from the disk subsystem into an intelligent storage node within the distribution layer, such as a storage router, which is also capable of replication functions. This disintermediation of storage functions provides customers additional choices of subsystems and could also facilitate easier integration or use of the IP core.

8.2.2 Designing an Effective Migration

An effective migration combines the maximum use of existing resources and personnel with the most easily implemented additions. IT professionals must consider a range of hard and soft costs associated with storage deployments that include the choice of devices, network fabrics , storage software, and personnel.

The flexibility of commoditized elements in the storage chain allows storage architects to combine the performance of specialized hardware platforms with the low-cost flexibility and functionality of intelligent storage nodes. In many cases, both types of platforms can be combined for an optimal solution.

One example of combining custom storage hardware with modular intelligent storage nodes is storage reclamation. This cost-effective , low-impact, high-return solution showcases the best of conventional and nonconventional approaches. Frequently, enterprise customers will need to complete major overhauls of their storage infrastructure at the expense of retiring older equipment. Consider, as one customer did, a forklift storage upgrade from SCSI-attached to Fibre Channel-attached disk arrays. The capacity and performance of the new Fibre Channel arrays significantly simplified the group 's storage operations ”so much so that they found themselves unable to find a use for the SCSI-attached arrays. The SCSI-attached storage could no longer fill enterprise needs in the data center. While it may have been able to suit less intensive departmental needs, the SCSI interface prevented easy access across a wide variety of departmental users.

Enter the intelligent storage node. Based on an x86 server platform, the customer made use of an iSCSI-to-parallel SCSI router with volume aggregation functions. This device "front-ended" the SCSI arrays and presented iSCSI-accessible volumes to departmental users over existing IP and Ethernet networks. The customer was able to effectively migrate this SCSI storage into a broader enterprise context, making use of an existing IP and Ethernet core with minimal additional investment in IP storage routers.

8.2.3 Calculating Total Cost of Ownership and Return on Investment

For years , total cost of ownership (TCO) and return on investment (ROI) have been used as measures for justifying IT investments. These quantitative measures attempt to boil down the myriad tangible and intangible costs into a simple, easy-to-understand "go" or "no go" decision on the IT project at hand.

The typical TCO calculation accounts for a range of cost factors, including the capital acquisition, installation and training, support, maintenance, and ongoing administration. The complementary ROI calculation measures the savings from completing the project and the payback on the initial investment over a period of one or more years.

Typical TCO and ROI analyses are project-centric and often do not incorporate synergies across enterprise IT expenditures. When planning for major IT investments, companies should consider the investment beyond the scope of the IT department. Treating IT as a corporate investment broadens the implications for business success or failure, and more accurately reflects the fit with company strategy.

Treating IT as corporate investment requires

• Demonstrated ROI path

• Short-, mid-, long- term company gain

• Ongoing match with business growth outlets

  The corporate planning process requires

• Consistency with company strategy

• Manageable cost

• Acceptable technical and commercial risk

• Competitive differentiation and sustainability

8.2.4 Managing an IT Portfolio

Beyond single IT projects, the true value of IT investments is realized through portfolio management. When considered in this framework, companies can take advantage of high-risk and low-risk projects with a variety of paybacks. Since many new technology initiatives fail, spreading investment dollars across numerous projects helps reap overall returns. The same principle has been applied with financial diversification for years.

Portfolio management is often confused with project management. The distinctions are outlined in Figure 8-6.

Figure 8-6. Separating portfolio management from project management. (Source: Mainstay Partners)

In the same manner that investment portfolio managers balance risk and returns, senior IT professionals must balance a variety of projects across the business criticality and innovation axis. A properly managed portfolio will invest in projects that fit within each of the sectors, as shown in Figure 8-7.

Figure 8-7. A technology investment portfolio. (Source: Mainstay Partners)