Chris Poelker

5 steps for evaluating your IT infrastructure

January 28, 2013 6:00 AM EST

In the pursuit of optimized IT models, chief information officers (CIOs) need to carefully evaluate their existing IT infrastructures and make choices that cut costs and streamline processes, as I discussed in my previous blog, Optimizing IT for 2013. To make that happen – and to deliver cost savings of up to 70 percent – following these five simple steps should make a major impact in improving your bottom line:

Step 1:  Dedupe your backup environment. Backup is the biggest headache facing most organizations. It is important for an IT department to take an inventory of the existing backup environment, business objectives for recovery time objectives (RTO) and recovery point objectives (RPO) for applications. In an optimized IT model, backup and recovery are typically the largest benefactors from a cost savings perspective. Dedupe to disk can save your organization a ton of money while still letting you leverage tape for long-term archives. Make sure the solution you choose integrates with any existing tape backup environment and enables you to automate tape for long-term archives AFTER data is replicated offsite. 

Step 2: Fix your network. The local area network (LAN) and wide area network (WAN) infrastructure must be robust enough to provide high performance for all client access while providing residual bandwidth for data movement between locations. Since WAN bandwidth is critical for data replication and in turn disaster recovery (DR), it must be robust enough to handle all production workloads while maintaining bandwidth to keep data in sync between sites to meet stringent RPOs. Since leasing WAN bandwidth can be the most expensive recurring cost for IT, it is critical to ensure that all data is optimized and minimal bandwidth is required. You can achieve an optimized IT model when you use global data deduplication and delta versioning at all levels so that when any data asset is moved, it moves as efficiently as physics allows. Invest in solutions that replicate data with maximum efficiency for DR and enable server-less, LAN-free backups to remove backup traffic from the network.

Step 3: Optimize servers and applications. In the world of open, distributed computing, server virtualization enables cost reductions through the ability to use less physical hardware for application services. This in turn enables more cost-effective DR, since fewer physical servers are required at remote sites. IT should run applications on robust “blade server” or GRID infrastructures or Domain- and LPAR-based physical servers, or use server-based virtualization solutions from vendors such as Microsoft, VMware and Citrix. Virtualization can put an organization ahead of its peers in terms of cost reduction and consolidation.

Step 4: Virtualize your storage area network (SAN). The SAN can be one of the most costly areas of IT, especially as it pertains to annual operational expenditure (OpEx), which can be four times more costly than capital expenditure (CapEx). Take all available measures to provide a robust, high-performing and reliable infrastructure for data storage in the most cost-efficient manner as possible. For example, an IT department can mirror data between two lower-cost modular storage arrays rather than purchasing high-cost monolithic storage arrays. The goal is to commoditize disk to keep costs down. An organization can move the intelligence for data protection and replication out of the storage arrays and into the fabric so that there is no need to purchase expensive, array-based software licenses. IT departments should be able to move data between unlike arrays at will while applications are up and running, and replicate data between unlike devices. Storage virtualization with fabric- or appliance-based intelligence makes disk a commodity and enables IT departments to keep vendors honest and costs low.

Step 5: Get your business continuity/disaster recovery (BC/DR) under control. Organizations benefit dramatically from an optimized IT model in terms of reduced BC/DR costs. Data deduplication and delta versioning can be used across all services to reduce storage and WAN costs. Low-cost iSCSI connectivity can be used in lieu of a SAN at the DR site. Fabric-level replication and virtualization can be used to achieve single console DR across all storage classes and vendors. This eliminates the need for advanced data service licenses within the storage arrays, and it enables replication between disparate storage resources. On-demand physical-to-virtual conversion and automated virtual-to-virtual recovery of application servers reduces downtime and streamlines the recovery process.

Implementing any of these steps individually will help reduce your costs, but together they create a compelling advantage in IT savings over the long run, and the New Year is a good excuse to start.