CTR Exclusives

Tips for switching to workload focused infrastructures

Tips for switching to workload focused infrastructures

By Joe Wolke

 

Today, businesses are facing greater change than ever before. Fortunately, they also have choices they’ve never had before… and the choices they make can and should be based on business objectives and cost optimization.

When the bridge in Minneapolis/St. Paul tragically collapsed, the country focused on its dangerously aging infrastructure; it seemed the time had come to really address this issue. But years later, many bridges with the same problems still stand, unimproved.  No other major collapses have occurred... But is it just dumb luck? Of course. The problem is just too big and expensive to easily solve.

Businesses do the same with IT. The cost to upgrade the existing model to meet current and future needs is so great that companies are doing what they can to eke another year out of hardware through low quality maintenance, or in some cases, a wish and a prayer.

Read more...
 

What’s Your Server Migration MPO and MTO?

What’s Your Server Migration MPO and MTO?

 

By Scott Van Dyke

The terminology recovery point objective (RPO) and recovery time objective (RTO) are commonly used to plan how to best protect data and applications around disaster recovery scenarios. For proper disaster planning, a business determines the maximum acceptable level of data loss following an unplanned event that could result in data loss (RPO). A responsible business will also determine the maximum amount of time they can tolerate without their critical applications (RTO). The primary focus RPO and RTO is to plan how to best protect your critical applications and data in the event of a disaster.

One often overlooked area that impacts a business is what happens to the critical applications and data for a business when performing a server migration. A server migration typically arises for one of the following reasons:

Read more...

Tiering: Scale Up? Scale Out? Do Both -- Part 2

Tiering: Scale Up? Scale Out? Do Both -- Part 2

 

By Mark Ferelli

Our exclusive interview with Hitachi Data Systems’ Hu Yoshida continues in part 2 of this feature. Click here to read Part 1.

MF/CTR: The more I hear it the more it sounds like the caching capability is at the foundation of the storage pool tiering infrastructure.

Hu/HDS: That’s right. It is the key because if you have those caches in separate nodes and you can’t use those caches as one global cache, then you’ve got silence. That’s where you split the workload between those silos to be able to utilize them and that’s more operational cost.

MF/CTR: If you have effective caching,  then you’ll be able to set up a foundation to both scale up and scale out without having to go crazy in terms of capacity?

Read more...

Oversubscribed InfiniBand Fabrics Cut Data Center Costs

Oversubscribed InfiniBand Fabrics Cut Data Center Costs


By Joseph Yaworski

Two factors that determine the cost and performance of a data center’s high-performance computing cluster (HPC) are the selection of the network interconnect technology and the design of the fabric. While many people view Ethernet as the least expensive approach—and InfiniBand® as the higher cost alternative—there are ways to design an InfiniBand fabric that can reduce the cost and provide better performance than 1-gigabit or 10-gigabit Ethernet (1GE or 10GE). Many applications benefit from the low latency of InfiniBand, but do not need the full bandwidth that it offers.  This means InfiniBand might be the best performing and most cost-effective network solution.

Read more...

Data Warehousing—Software as a Service?

Data Warehousing—Software as a Service?

 

By John K. Thompson

The advances we regularly record in technology make more opportunities available to us almost daily, it seems.  Those opportunities aren’t always apparent to those making the initial advances possible.

For instance, companies struggled for years with the question of which data to keep in their IT infrastructure after it was no longer immediately needed.  The cost of storage was prohibitive, of course; keeping gigabytes of data was not seen as economically feasible, especially in an online implementation.  So, data was regularly trashed, or moved offline onto tape, which was pretty expensive as well on a per-megabyte basis.

Read more...
Page 6 of 26