Call: 469-635-6200
Name *
Phone No. *
Email
Question
Security Code *
Refresh
 
Submit
Name *
Email *
Phone
Question *
Security Code *
Refresh
 
Submit

Get more from your data

Scalability Experts Blog

Private Cloud Reference Architecture

 

Private Cloud Reference Architecture

Recently, Microsoft released a new and improved Fast Track Reference Architecture for the Private Cloud.  Why is this important?  As technology changes, often best practices and guidance does as well.  With the impending release of Windows Server 2012 and the capabilities baked into the product, this new reference architecture can help customers take advantage of the private cloud.  One of the best things about a reference architecture is that it is a reference.  Microsoft took a lot of time to document guidance and best practices, but they did not call out specific hardware.  That is understandable because Microsoft has partnerships with many manufacturers.

There is a paragraph about hardware that is generic.  This is where I want to drill down a little deeper.  We will start with the smallest form factor, generally a single width 2 processor blade system.  The big advantage is simplicity and price.  These small blades may be clustered together in a Hyper-V cluster resulting in the potential of a lot of nodes within a blade enclosure.  The main disadvantage is that the small form factor also limits RAM and space for additional HBA or Network cards.  It also limits the parameters of VMs in terms of vCPU and RAM.  There are also double-width blades.  As you would imagine, these blades have more processors (up to 4), RAM, and space for expansion cards.  The limitations on VMs aren’t as severe in terms of computing power because some of these blades may have upwards of 40 physical cores; however, RAM is still somewhat constrained.  Generally, blades are a very cost effective way to accommodate the need for computing resources; however, they also lack space for additional cards.  Depending on a number of factors, blades may be sufficient for building a Private Cloud for databases.  While blades come in two sizes (single and double width), rack mounted systems usually come in 1U, 2U, and 4U sizes.  Typically the larger the form factor indicates more processors in the system. The larger form factor systems also mean that the systems can hold more RAM and have more slots available for HBA and Network cards.

Why am I so focused on Space?  IO.  The data for a database usually lives on some storage device. Even though the storage device may be sufficient in terms of speed, a system may not have a big enough pipe to take advantage.  This is often remedied by adding more cards to build a bigger pipe to the data storage device.  Having space available to build that bigger pipe may make the difference in implementing a successful Private Cloud Infrastructure fit for hosting databases.

Will blades suffice for your organization?  Or do you need rack mounted servers?  Often this is a matter of preference.  But you do need to know your current IO demands of the workloads that will be moving to the Private Cloud Infrastructure and allow for growth as well.  Scalability Experts is ready to help you determine the optimal solution for your organization.  Take advantage of SE’s experience with multiple environments and start your journey to the Private Cloud the right way.


-Authored by Rick Heiges, SQL MVP and Solutions Architect with Scalability Experts

 

comments powered by Disqus