Budget Time and Resources to Manage your Infrastructure Lifecycle

9 minute read
Managing the Lifecycle of Infrastructure

"Don't put off until tomorrow, what you can do today." – Ben Franklin

It is no secret that maintaining the lifecycle of your infrastructure can be one of the most boring jobs in information technology (IT), but it also happens to be one of the most important.

It could be tempting to let tasks related to maintaining infrastructure fall to the wayside, but don’t fall into this cycle of procrastination. Maintenance tasks are just as important as software development to keep your organization running. Think of the infrastructure maintenance cycle as the time leading up to your final exams in college. Is your best bet to wait until the last possible minute and pull an all-nighter, or to study a bit each day to understand and absorb the material?

It’s important that you get to know your infrastructure and be consistently mindful of its maintenance. To properly address infrastructure lifecycle management, ask questions about your environment, even if they lead to uncomfortable answers.

For example: Are you still using the Windows XP or Server 2003 platforms? Both of these have been discontinued and no longer have extended support from Microsoft.

If you have to answer yes, don’t worry – you’re not alone. The governments in the United Kingdom and the Netherlands, for example, have each paid Microsoft millions of Euros for ongoing support of the now-defunct Windows XP.

The end of “extended support,” in Microsoft-speak, means no more security fixes, updates or bug fixes.

Let’s repeat that: No more security fixes.

That is a major red flag and a critical reason to avoid procrastinating infrastructure lifecycle management.

Server 2008 has reached the end of mainstream support (as of January 2015), which means organizations that are still using it, only get self-service or paid support options. While Server 2008 will remain in extended support until January 2020, it’s not a recommended practice to wait until then to come up with a transition plan. You could easily end up in the same situation as Windows XP or Server 2003 users.

This issue isn’t limited to Microsoft, of course. All software vendors need to remain patched just to address vulnerabilities and compliance issues over time. Don’t forget the firmware and BIOS versions either.

We’re Not Just Talking about Operating Systems

Firmware updates are not just for initial deployment. Hardware manufacturers create and push out their own download processes and tools to facilitate updates, but one challenge is scheduling the necessary downtime to follow these instructions. For example, bare metal servers, such as a Windows server running on a standalone rack mount server, will require an outage--one of the major reasons that all servers should be virtualized if they can be. While servers running a hypervisor can usually migrate their virtual machines (VMs) to perform server maintenance, it is all dependent on the cluster design. It relies heavily on whether or not you have maintained absolute vigilance when it comes to the capacity planning of the environment.

Don’t forget that inter-compatibility matrices need to be maintained also. For each software or hardware manufacturer, there is a very specific list of tested and supported versions that need to be maintained in order to solicit support. The first step of troubleshooting any issue is almost always to validate version and compatibility.

But, you can’t just look at the servers. You need to look at your upstream switches, your storage area network (SAN) fabric and arrays, and of course the applications that will sit on top of this infrastructure. It all has to play well in the sand box together.

Don’t Let Your Infrastructure Down

Although every environment is different and unique, there are general rules that can be followed to ensure you’re keeping up with your infrastructure management.

1.    Understand what you have

Remember that exam you have to study for? Isn't it easier to understand what may be on the exam if you've been studying throughout the entire semester? You can't open a textbook for the first time when you sit down for an all-nighter and expect to have a good understanding of the material.

Begin by understanding what you have and how it works together. A configuration management database (CMDB) is a critical tool to employ for this task. Whether it's a simple spreadsheet or an expensive, complex solution, it must hold all of the configuration items (an "entity" that has specific configurable attributes, such as a physical device or server), logical devices (such as a database instance), or conceptual devices (business process).

To aid with infrastructure management, your CMDB needs to have all of the physical servers and their specific hardware (NICs, HBAs) and firmware / BIOS versions. Assuming most enterprises are virtualizing their data centers, all of the VMs should be documented as well as what cluster / datacenter they reside in. We know that we can't document the physical server they reside on, just for the fact of vMotion, right? Lastly, document the applications that reside on each VM, who the business owner is, and the specific software and version.

2.    Know the time commitment and plan ahead

If you plan to budget study time for a big exam, you begin by figuring out the information you are responsible for, and then you create a plan for learning it. Similarly, if you manage any type of infrastructure, you know that new versions, patches, and fixes must be factored into your workflow in order to keep up with infrastructure lifecycle. It’s important to budget time for infrastructure maintenance checks at the correct intervals so that these tasks don’t fall to the wayside. If you fail to plan for your lifecycle management, you really plan to fail. This could result in spending entirely too much money to remediate issues within your infrastructure in an expedited manner, or even failing to meet deadlines which could have dire consequences when there are underlying issues, such as compliance.

Your physical and virtual operating systems are a good place to start: You can begin to understand the immediate needs, and then plan for the next two to five years.

For your reference, below are excerpts of the VMware and Microsoft Server support matrices. Comparing your inventory list to the charts below would be a good starting point to get a rough idea of the status of your infrastructure.
 

Product

Lifecycle Start

Mainstream Support End Date

Extended Support End Date

NT Server 4.0 *

7/1997-11/1999

12/31/2002

12/31/2004

Server 2000 (all versions)*

3/2000-5/2001

6/30/2005

7/13/2010

Server 2003 R2 (all versions)

3/2006 – 3/2007

7/13/2010

7/14/2015

Server 2008 (all versions)

5/2008-5/2009

1/13/2015

1/14/2020

Server 2008 R2 (all versions)

10/2009-2/2011

1/13/2015

1/14/2020

Server 2012 (all versions)

10/2012-1/2013

1/9/2018

1/10/2023

Server 2012 R2 (all versions)

11/2013

1/9/2018

1/10/2023

* NT and 2000 versions included for reference only. It is our hope that all customers have migrated off of these un-secure, unsupported platforms years ago.

 

Product

General Availability

General Support

End Date

Technical Guidance End Date

ESXi 3.5*

1/10/2008

5/21/2010

5/21/2015

ESXi 4.X

5/21/2009

5/21/2014

5/21/2016

ESXi 5.0 and 5.1

8/24/2011

8/24/2018

8/24/2018

ESXi 5.5

9/19/2013

9/19/2018

9/19/2020

ESXi 6.0

3/12/2015

3/12/2020

3/12/2022

* ESXI 3.5 reached the end of extended support 5/21/2013.

 
3.    Have a strategy

A lab environment is a must-have with the frequency that new hardware and software is released. A lab or development environment should mimic production as much as possible. It should be a place that can be used to test new generation releases of hardware (not limited to just the box, but HBAs, NICs, etc.), storage, network components, etc.

The lab is where you get to “write and erase” your material in order to review it over and over again until you understand it inside and out. Think of it as your flash cards, study guides, and white board rolled into one study tool.

The lab serves as a place to test hardware and operating system (OS) compatibility, but it should also be used to test processes. Examples include upgrading your active directory schema in preparation to move to Server 2012 AD, moving off of Exchange and Office to Office 365, and establishing processes for guest applications. Most compliance projects require a written and tested process prior to the actual change, so this would help satisfy that requirement.

Development time in the lab is also a good place to test out new deployment tools or methods, such as automation and orchestration. Much of the software can be downloaded with temporary or evaluation licenses. Some vendors may even allow you to upgrade the licenses for a nominal fee.

4.    Include the business/application owners in your planning

Just as tutors and teaching assistants can help review curriculum and develop study guides, collaborate with your line of business leaders to make strategies for change.

Make a plan with the business or application owners well before a change is needed. It’s important for both sides to understand that some applications are no longer supported at all, are not supported on certain versions of hardware/software, or require special provisions for the change (like serial dongles plugged into the back of a server). It is important to include the stakeholders in discussions about how you will prepare to move or upgrade to new infrastructure or operating systems.

5.    Don't forget your partners

You might grab the math whiz to help you with your differential equations class, and the pre-med student to help you with your biology lab. You don't have to go it alone, you just have to know where to look for resources.

Some vendors may only want to sell you the latest and greatest software or hardware, but your trusted partners want to help you to succeed with the right products and services. Don't forget to lean on them to help with this process. It can be daunting if you are starting out fresh.

A few tasks that your partners should be able to help you with include:

  • Hardware/software compatibility and interoperability (if all items are purchased on the same bill of materials)

  • Deployment services to ensure everything is working together and patched appropriately

  • Flexible services such as first-call support, staff augmentation or even assistance performing upgrades during a change window

When can you start?

So, are you doing everything you can to ace that mid-term or final exam? Have you been studying the whole semester and seeking out the right resources for help? Or have you been procrastinating and finding yourself struggling to understand the coursework? Don't make it harder than it has to be. Create a plan for preparation now, so you’ll be successful in the future.

For example, with Server 2012 R2 released in November of 2013, isn’t it time that you came up with a migration strategy? Have you talked to your application owners to talk about migrating their apps to a new server platform? Have you upgraded your active directory yet? Have you even tested out the migrations in a lab environment?

Whether we call it patch management, version control, or lifecycle management, these tasks must be factored into your primary IT responsibilities and budgeted out correctly throughout the year.

With vSphere 6 released to the public now and Microsoft Server releasing their next version in a short year, we should all be testing these products in the lab and talking to our software vendors about supportability and updates. And of course, we should be accounting for this type of work in our budgets next year, and every year following that.

When it comes to infrastructure lifecycle management, you have to choose between having a reactive or proactive organization. Prepare yourself throughout the year so that you don’t get caught suddenly scrambling to “pass the final exam.” Budget your resources each year; make an inclusive plan; take initiative and collaborate alongside other business units. With the right approach, an infrastructure lifecycle strategy isn’t a dull requirement of an IT organization, it’s an opportunity to be a true partner to the business. 

Transform your business with cloud
Leave a Comment

You Might Also Like

About the Author

Popular Today

Slideshows

Videos

@ForsytheTech