The Cloud Scoop from Fedscoop

In this article, eWeek provides a good summary of a Fedscoop cloud computing panel event.  Below, are some of the “soundbytes” accompanied by my commentary.   

 

According to the experts, some benefits of Cloud Computing are:

1.      “lower capital expenses “

2.      “it’s scalable”

3.      “it’s real”

4.       “it’s a high-performance option for most agencies”

Certainly #1 and #2 are valid.  #3 depends on the particular cloud services (e.g., SaaS, PaaS or IaaS) and industry.  There is definitely hype in some vendor offerings.  I’ve already offered a laundry list of cloud computing benefits

Critical considerations for cloud computing:

1.      “is not cost, it’s flexibility”

2.      “Microsoft refers to that flexibility as agility”

With regards to #1, this sounds similar to an open source movement’s battle cry.  Cost definitely needs to be part of the migration equation.  Flexibility alone isn’t enough justification unless the costs of each solution are near equal.  Of course, I do agree that one needs to evaluate the total cost (knowledge transition costs, existing systems data migration, personnel costs associated with the migration, etc…) associated with the migration in the analysis.

With regards to #2, I’ve already commented on the distinction between flexibility and agility. 

 

 

Where to start in the Cloud: My Commentary:

1.    All new applications

Definitely.  Any new investment should be considered a candidate for the cloud

2.    Software license expiration

Absolutely.  You are probably paying for more asset than you actually use.  And, the vendor might already have a cloud-based service offering. 

3.   Collaboration apps

Certainly.  Collaboration apps, such as email, are a commodity and should be jettisoned from the control of corporate IT.

4.    HR

Probably.  Like commodity apps, HR is a common function across enterprises.  Although some HR functions (e.g., talent management) and associated software automation and analysis could be considered strategic and thus not easily replicated using existing commercial SaaS.

5.   Dev and Test

Yes and Yes.  Although one needs to consider potential performance implications as some development is computationally intensive (i.e., need excellent bandwidth) as well as computation prowess.

 

Where not to start in the Cloud:

§  ERP

§  Transaction processing

 

Yes, agreed on both categories.   And, obviously I would add any data-sensitive application.  The primary reason for this is that cloud-based assets don’t have the same legal protection standards that on-premise, enterprise owned and maintained systems do.

 

Cloud Pain Points:

§  Integration of compute, network and storage infrastructure

§  The issue of VM (virtual machine) sprawl

§  Need for more and better management tools; security

§  Federation of the cloud

 

Yes, integration of cloud-to-cloud and cloud-to-on-premise systems is indeed a challenge.  As is the overall management and governance of such a federated system. 

Posted By : Ray Bordogna

Leave a Reply