Media Center / iWave Blog

Storm Front: Hot Cloud Tech vs. the Cold Reality of Existing Infrastructure

By Brent Rhymes

Sitting in the Alumni Lounge at EMC World 2011 today, I overheard a conversation between an EMC partner and customer in which the word "cloud" was used about four times in every sentence. Adjective, adverb, noun, verb... didn't matter...  
 
It reminded me of the lyrics from the 1960's hit song "Surf'n Bird, A-well-a everybody's heard about the Cloud, C-c-c-cloud, cloud, cloud, c-cloud's the word, C-c-c-cloud, cloud, cloud, c-cloud is the word."  
 
While even the EMC World booth talent get the fact that "the Cloud" is a transformational event for our industry, the vendors pushing the "journey to the cloud" prefer to ignore earthly matters such as the investments enterprise IT organizations have made in their existing IT infrastructures and "ITIL processes" for problem, incident, change, configuration and event management.  
 
This issue of marrying the new (cloud) with the old (existing IT infrastructure) appears to be an especially big challenge for the *Cloudiest of the Cloud Vendors," those peddling "Converged Infrastructures."  
 
Converged Infrastructure vendors combine network, compute and storage resources into a "pod" or "block" that is designed to host virtualized environments (aka, white fluffy clouds). The two "big" ideas behind pod/blocks are (1) keeping a tight, prescribed infrastructure will help reduce operating costs and (2) pod/blocks are scalable - just add more pod/blocks to your deployment as the resource demands of your cloud grows. For a much better definition of "Converged Infrastructures," see http://en.wikipedia.org/wiki/Converged_Infrastructure . Vendors offering Converged Infrastructures include HP(FlexNetwork), NetApp (FlexPod) and the VCE (vBlocks). There are more, many more, in this space... Much more on "converged infrastructures" in a later blog post.  
 
In the early days, Converged Infrastructure vendors dreamed of "green field" implementations for their pod/blocks (why they didn't use "blue skies" as the metaphor is beyond me). In this fantasy, large organizations would build new cloud infrastructures using the pod/blocks that had little to no dependencies on the organizations existing IT infrastructure.  
 
Over the past 12-months, reality has rained down... and enterprise IT customers have spoken... and it turns out that they would like to leverage their existing infrastructure to manage their most strategic projects (cloud-based initiatives):  
 
- They want to manage pod/block infrastructures with their established tools for asset, change, event, incident and problem management. They want an approved change request in their service desk to automatically effect changes within their pod/blocks.  
 
- They want a central, audited and unified process of provisioning both additional pod/block hardware as well as provision the VMs that resides on the pod/blocks.  
 
- Enterprise IT organizations abhor vendor lock. Most want to implement cloud infrastructures that enable them to substitute vendors for the network, compute and storage - and many want the flexibility to substitute hypervisors.  
 
- They expect the pod/block vendors to adapt to their systems for managing IP addresses, configuring security, registering new machines, etc.  
 
- They need a unified platform to automate their procedures for disaster recovery, workload management, fault remediation, etc.  
 
So how are the Converged Infrastructure vendors dealing with these challenges? Are they packaging IT automation and orchestration solutions with their pod/blocks? How do they balance intra vs. inter pod/block orchestration?  
 
All excellent topics for my next post on the differences between enterprise vs. embedded orchestration for Cloud Service Management.