Michael Jude, Compute Platforms Manager, Alaska Airlines
Alaska Airlines is dedicated to “Creating an Airline People Love” and in IT, we support this business mission by providing innovative technology solutions to our employees, as well as to our 44 million guests worldwide. Like many companies today, Alaska Airlines has embraced the public cloud and has initiatives in place to expand our footprint in this space. However, existing investments in world-class datacenters as well as the time it takes to refactor legacy applications into modern cloud architectures means that our datacenters will be around for the foreseeable future. In order to keep up with business demands and shifting expectations of our business partners due to the presence of public cloud, IT and specifically infrastructure teams need to evolve to meet this demand.
Traditionally, infrastructure has focused on providing reliable and performant systems at a cost-effective price. In today’s ever-changing environment those requirements are as important as ever but additional expectations around delivery speed, availability of resources, and frictionless deployments are pressuring infrastructure teams to be more cloud-like.
At Alaska Airlines, we have embraced this challenge and have implemented innovative solutions using off-the-shelf VMware products to provide detailed analytics of our environments, increasing reliability and performance, as well as creating a 100 percent hands-free automation solution for ESXi host provisioning. Implementing these new solutions and processes facilitates our ability to create a private cloud that is adaptable to changing business needs as well as providing the flexible infrastructure that our application partners require. Our end goal is to provide cloud-like capabilities so our application teams are not concerned about which infrastructure they run on—our hybrid cloud powered by VMware will provide the features and benefits they need to keep pace with our business.
Alaska Airlines, with contributions from Cisco and VMware, have developed a fully automated solution to auto-scale on-premise virtualized infrastructure so our private cloud can provide the necessary resources as needed. In early 2018, an additional 30 percent of capacity was required across multiple environments to support the recent acquisition and integration of Virgin America. Using our automation solution for host provisioning we were able to add this capacity in one day; Our previous build process would have taken us weeks to accomplish this goal. More importantly, because these are built using automation we can sleep at night knowing that they are configured the same way every time — removing human error and eliminating configuration drift.
The focus of our solution is to increase the deployment speed and reliability of VMware ESXi host provisioning. We accomplished this by leveraging Cisco Unified Compute System (UCS), the VMware hypervisor (ESXi), VMware vRealize Operations (vROPS) and other VMware products.
The combination of these tools created a Software Defined Datacenter (SDDC) that based upon capacity alerts from vROPS, would trigger workflow automation to build a ESXi hosts. The workflow utilizes rules, policies, pools and configuration items in an orchestrated process utilizing our available tools.
Our VMware powered hybrid cloud, Alaska Airlines is eliminating traditional infrastructure challenges and enabling a future of seamless workload mobility between any cloud
Before moving on, we need to mention the importance of vROPS for highly virtualized environments. The critical insights and performance data it provides about virtual infrastructures is vital to help identify issues, provide capacity planning metrics, and enable proactive monitoring; without a tool like vROPS, managing a private cloud would be difficult at scale. More importantly, by using empirical performance data, infrastructure teams can partner with application owners to optimize their environments — ensuring that their applications are performing efficiently.
Portable workloads to any cloud
Alaska Airlines has a multi-datacenter and multi-cloud vision enabling a highly available, reliable and cost-effective hybrid cloud. This vision challenges not only our infrastructure to be portable to any cloud but also other supporting technologies like network security, Disaster Recovery, backups, and application delivery. By utilizing VMware’s hypervisor and vRealize software suite, we have solved traditional infrastructure pain points using automation to create a more reliable and scalable private cloud. But how do we transition these workloads to the public cloud, if necessary? Fortunately, VMware has an answer for that too.
VMware recently announced a new technology called Hybrid Cloud Extension (HCX) which promises to accelerate hybrid cloud adoption. HCX enables hybrid cloud designs by providing an interconnect between compatible vSphere environments. The solution extends layer 2 networks eliminating the need to re-IP or re-factor existing workloads, provides WAN optimized live vMotions or bulk VM migrations depending upon the use case. Additionally, HCX can be used as a Disaster Recovery as a Service (DRaaS) service to quickly shift workloads from a degraded site to a healthy site — a pain point with traditional infrastructure.
We are just touching the surface on what HCX can provide for workload portability between multiple public and private clouds, but the future is bright. With our VMware powered hybrid cloud, Alaska Airlines is eliminating traditional infrastructure challenges and enabling a future of seamless workload mobility between any cloud.
Datacenter services at ROBO
Alaska Airlines operates out of more than 140 airports across North America, Mexico, and Central America, and each of these remote locations requires local IT services to support our operation. Traditionally, infrastructure has provided a standalone, physical server to provide the multiple functions required to operate efficiently, which can cause conflicts during upgrades, failure events, etc. Over the past 12 months, infrastructure engineers have been replacing standalone servers with ESXi hosts to provide the ability to split services across multiple VMs removing the previous challenges. However, at our most critical locations, we needed a way to provide datacenter like resiliency at a cost-effective price.
Virtualizing Remote Office/Branch Office (ROBO) servers has been a trend for years, but there hasn’t been a great, cost-effective solution to provide resilient infrastructure without investing in a local SAN or relying on a new vendor … until now. Released with vSAN 6.1, VMware introduced a 2-Node vSAN with Remote Witness which satisfies our needs for our most critical remote locations.
Today, Alaska Airlines has deployed our 2-Node vSAN with Remote Witness solution at multiple critical remote offices with great results. With this solution, we have surpassed business expectations by deploying a corporate wide CDN solution enabling greater communications and transparency from leadership as well as proving that this solution is stable enough to support Tier-1 workloads, which will be going live at multiple locations later this year.
By continuing to adopt new VMware technologies and creating innovative automation solutions, Alaska Airlines has improved our hybrid cloud approach and we are well on our way to making our multi-datacenter/multi-cloud vision a reality. Through our strategic partnership with VMware, we are creating a fully elastic and functional cloud that enables our teams to surpass business expectations and ultimately provide the best experience to our customers — In IT we are passionate about creating an Airline People Love by connecting people through technology, our guests to our employees, our employees to our business, and our business to our Community while continuously innovating at the speed of flight.