Cloud Computing for System Administrators
The job of a system administrator is changing in the face of cloud computing. As systems move to the cloud, attention shifts from managing physical resources to managing virtual systems. Organizations will look to system administrators for leadership in architecting and rolling out cloud environments. It can feel like a daunting task with so many cloud vendors available. Instead of focusing on the vendors and cloud technology, let's examine what this means for system administrators as they tackle these new initiatives.
For system administrators working for organizations with an existing physical server presence, the number one step is to migrate those systems to the cloud.
There are many concerns and issues that can arise depending on the vendor or vendors chosen.
Does it make sense to rebuild the entire system from scratch?
Does an existing system image that offers 80%+ of the functionality needed? The most reasonable solution may be to make a virtual copy to migrate to the cloud vendor. Sysadmins need to understand the tools and support options available to them for moving systems to the cloud.
Organizations with a large amount of data, e.g. hundreds of gigabytes or terabytes, must have a strategy for moving that data to the cloud.
Sysadmins can develop a migration plan that is dependent on the frequency of access and change to the data.
Other aspects come into play.
An organization could mandate that no data becomes inaccessible at any time during the migration. These constraints can affect the level of effort necessary. System administrators will need to provided their expertise to level expectations with costs and time.
Thankfully, vendors offer numerous options while many third-party tools, both commercial and open-source, are available. All of these options can significantly save on time and costs if implemented correctly.
Integrating with existing environments is almost unavoidable when moving to the cloud.
Local networks often need to connect with the cloud environment.
This requires provisioning of both local and cloud resources with compatible protocols.
If the sysadmin is combined with the network administrator role, establishing the proper connections through route tables will be necessary. Knowing how to configure resources to make and keep a successful connection then becomes an essential skill.
Some organizations may want to bridge together two or more cloud environments from different vendors.
This is certainly possible with the proper design.
Sysadmins can design and build hybrid environments that offer the best in terms of performance and cost to meet requirements. The challenge for them is understanding which vendor is best suited for each specific need of the organization. Innovation by cloud vendors means sysadmins need to stay up-to-date on the various offerings.
With the advent of cloud computing, system administrators finally have renewable, recyclable resources that are easy to provision.
Creating a new server is as simple as a few clicks.
Replacing an existing server is just as simple.
On top of the ease, provisioning takes far less time in the cloud than in traditional environments.
Made a mistake in choosing server specs? Image the contents and start up a new server with that machine image.
Managing a cloud environment means more than just keeping all systems operational.
Administrators must now keep an eye on the cost of running the environment.
Hardware costs have been replaced with paying per compute hour, sometimes with licensing costs included and sometimes not. Warranties are gone.
Support contracts have moved from the hardware vendors to the infrastructure vendors. Managed services trade full control over the setup, maintenance, and general operations with lower cost, pre-built common scenarios, and little to no required maintenance. Understanding these trade-offs will help sysadmins work with others to build a cost-effective, reliable solution.
Security is a major concern in the cloud.
Everybody involved in the cloud environment has a responsibility to secure it.
Vendors are responsible for securing physical access to the hardware.
Securing the network and other provisioned resources falls on the organization. In smaller organizations, this could mean the sysadmin owns it.
Whether in charge of the security or not, the admin should be familiar with his or her cloud vendor's security model.
Roles, groups, users, and policies are common mechanisms for granting and restricting access.
Most cloud vendors have created granular permissions.
For example, a policy might grant permission for a user to create an instance in one region and deny it in another. In cloud environments, this can be an issue.
A sysadmin will have a lot of work to do If a developer accidentally creates a script that spins up resources and fails to shut them down. Therefore, prevention is key. Just like traditional environments, security should be an upfront consideration.
Permitting or blocking network access, a role traditionally owned by network administrators, is something easily configured in the cloud.
System administrators are better off understanding how this is done as well as when to use a particular method of securing resources over another.
This is especially true given how cloud-based jobs tend to merge together because of the convenience. Plus, the more skilled an administrator is in managing the environment, the more in demand he or she will be.
Once in the cloud, an organization will expect systems to be available at all times. This does not happen without some careful consideration. Thankfully there are many tools at a sysadmin's disposal to meet uptime requirements.
Cloud vendors do a very good job of integrating their datacenters usually based on proximity to one another.
This makes regional infrastructure capable of withstanding single datacenter outages if implemented to handle that situation.
A sysadmin should know which managed services offered by his or her cloud vendor support multiple datacenters and how those managed services can be utilized in the organization's environment. Examples include running a managed database with failover capabilities, websites distributed across multiple datacenters, and remote storage.
System administrators cannot be expected to work 24/7 to guarantee uptime.
As a result, the use of automation becomes a critical component.
Without it, scaling applications might not happen when systems really need it.
Sysadmins can use various metrics, and even custom metrics, to detect events. Those events can trigger scaling activities and notifications such as email and text messages. Configuring the right combination of event responders can make a sysadmin's job much, much easier.
Headaches caused by hardware failures do not necessarily go away in the cloud but the chances of encountering them are significantly reduced.
Underlying hardware failures do happen and typically can be recovered from assuming the proper steps were taken to backup data.
System administrators familiar with methods for backing up systems and data can establish realistic recovery time (RTO) and recovery point (RPO) objectives. In the cloud, an admin can provision environments to validate RTO and RPO without affecting production environments.
This is useful for auditing plans especially for 3rd party audits that expect recovery plans to be met.
A common reluctance for moving to the cloud is outages that are beyond the control of the sysadmin.
Instead of being caught in that situation, sysadmins can prepare for the inevitable by running infrastructure globally.
If a single region becomes unavailable, another region can be made active. Treating infrastructure as code gives admins the ability to bring up new regions with relative ease. In the rare event of a global outage for a particular vendor, the infrastructure code can be used as a resource for building an environment with a different vendor. This would require a system administrator with multi-vendor cloud knowledge and experience.
The disaster recovery approach is influenced by other factors such as the budget and how important it is to keep the environment running with maximum uptime.
Maybe the best approach is to recover to a traditional datacenter environment yet that options is cost prohibitive. The sysadmin will be the expert in meeting these requirements. Without a strong understanding, he or she runs the risk of spending too much, not providing for an accurate recovery, or both.
At Cloud Academy, we have developed learning paths to help system administrators become top notch cloud architects.
Whether you plan to build cloud environments or work for cloud vendors, knowing the cloud is critical to your growth. Cloud Academy is the “go.to” resource for knowledge on cloud vendors and technologies. Start your learning path today.