ESG Blogs

We often hear small and mid-sized organizations talking about how management constantly urges them to “do more with less.” Keeping competitive requires them to squeeze all they can out of their infrastructure, and their staff. ESG research bears this out; respondents chose cost reduction as one of the business initiatives that will drive the most technology spending at their organizations this year, appearing second on the list, only bested by strengthening cybersecurity.

Let's face it, you built your data center for one reason: to generate revenue. The return on investment of your hardware, software, and people really depends on what you do with that infrastructure: Is it mostly used for production activities? Or is there a lot of time when users are stymied because there’s a backup running? How much time is spent on essential—but non-production—activities like data protection?

Backup is the first line of defense, and if you’re like most organizations, you do some kind of backup daily, weekly, and monthly. To get the most out of your infrastructure and IT staff, you want to spend as little time and effort as possible on backup, but still be ready to restore fast. To that end, a common practice is to do incremental backups daily, and then once a week or so roll those incremental changes up into a synthetic full backup. This gives you a full backup that’s easier to restore from, but taking less time out of production. 

EMC’s early challenges included long provisioning times, low storage and CPU utilization, not enough data center space or power, and an IT staff overwhelmed by growing demand. Sound familiar?

A lot of their success is summed up in the graphic below, which shows progress from 2004 and continuing into the future. I’ll stick to the highlights : EMC saved $66M in OPEX and $157M in CAPEX, reduced CO2 emissions by 100M lbs, and has increased both productivity and agility.

IT has a complicated relationship with redundancy. Redundancy is built into equipment and processes as a hedge against failures – that’s good. We put redundant power supplies in our equipment so that if one fails the other takes over. We use RAID in various permutations to ensure data protection. We put up duplicate data centers so that if a primary site gets flooded out we have another location from which to work. And we do backups every day and/or every week to make sure we have a redundant copy of our applications and data.

And that’s where it gets complicated, because backing up the same data over and over again, day after day, week after week, has a big impact on how much storage we need, how much bandwidth we need for backups, and how long it takes to restore. Redundancy is good – but too much of a good thing can become a problem, affecting cost and performance.

There used to be a clear distinction between IT infrastructure components and an obvious division of labor – the application, server, network, and storage teams operated mostly independently. End-users and business managers were content to follow IT’s instructions, being told what applications were available and then figuring out how to leverage them.

In recent years, however, much has changed. Application virtualization, management automation, increasing “consumerization,” and the concept of service-oriented IT have conspired to dramatically increase user expectations and demands. At the same time, IT is now viewed more as a strategic asset, an arrow in the quiver that the business can use to hit customer and employee satisfaction targets. Instead of waiting to be told by IT what capabilities are available, constituents approach IT to request services, applications, and results.

At VMworld 2011, VMware announced that they had achieved a milestone - 1 million IOPS on a single vSphere 5 host with multiple VMs. To do that they used a traditional, spinning disk storage array with eight engines and 980 15K RPM Fibre Channel disks.

Does it make sense to run parallel storage infrastructures for all the tasks you need to do? A lot of IT managers are doing that – one infrastructure stack for physical servers, another for VMs, SAN for some applications and NAS for others, still more silos for backup, DR, etc. What a headache. And virtualization, while offering great benefits, often turns out to be the straw that breaks the camel's back.