Last month, Rackspace hit a major milestone, surpassing one billion hours of OpenStack under management.
During that time, we’ve learned a lot of lessons — not only from a technology standpoint, but also about what helps our customers succeed. Meanwhile, those customers’ use cases are becoming increasing complex, with some achieving a faster rate of technology maturity than others.
Yet one use case present in nearly every industry is multi-cloud.
Multi-cloud can be defined in different ways. For example, moving a public cloud to another public cloud provider could be considered a multi-cloud approach. A private cloud interconnected to another private cloud is also considered multi-cloud. For this post, we’ll focus on integrating private cloud to a public could provider.
In order to understand why an enterprise would choose to pursue a multi-cloud strategy one must first understand the four fundamental reasons to use a private cloud. Some organizations may fall into one or more of these categories.
- Compliance: if your organization is part of a regulated industry, such as medical, retail or financial services, a private cloud is purpose-built, and so designed to specific industry and regulatory standards.
- Security: for some organizations, the risk of having their data in a public cloud make it a no go. Not that a public cloud can’t be secure, but the sensitive nature of some business data may require (or bring added comfort) that all data reside on-premises.
- Locality: public clouds are not always located in regions advantageous to a business. For example, if your organization delivers or consumes low-latency services that measure success in milliseconds, private cloud is the only platform that will deliver a consistent, deterministic experience.
- Performance: one of the biggest benefits of a private cloud is not having to worry about “noisy neighbors,” since the pool of resources are dedicated and not shared across multiple cloud tenants. This means all your workloads will effectively perform in a predictable manner that will be easy to baseline.
Another point worth mentioning is that Rackspace Private Cloud yields the benefit of a static operational spend, because all resources are dedicated to a single tenant, regardless of use. This is a shift from public cloud utility-based billing, but in most cases, our customers see dramatic cost savings with Rackspace Private Cloud.
Now that we’ve covered the main reasons an organization would opt for private cloud, let’s drill into some of the potential use-cases for multi-cloud strategy.
Historically, one of the drawbacks of private cloud has been the inability to burst beyond your allocated resources. I.e., if an organization has “x” amount of compute resources available to consume, proper capacity planning is required in order ensure the compute pool has additional resources during peak demand.
At Rackspace we perform capacity planning for our private cloud customers and we’ve found that it’s easy to predict scaling — if demands are linear. But when a private cloud needs to scale in a non-linear manner due to an unplanned event (think natural disaster, market shift, or something as simple as a non-performant code promotion) that requires additional resources on the fly.
This is generally the most compelling use-case for multi-cloud.
Organizations often say they’d like “workload portability” — the capability to move an application instance from private to public and then perhaps back to a private cloud. That sounds like cloud nirvana, and in theory, the technology exists today to make it happen — but the cloud industry is not quite there yet. We still need to deal with latency between the private cloud and the public cloud provider as well as network interoperability and peering.
How are we addressing these challenges at Rackspace?
Rackspace Private Cloud, is Rackspace’s distribution of OpenStack, which we offer as a managed service. Our OpenStack private cloud can be deployed within a Rackspace datacenter, or on-premise within a customer’s data center.
During the Austin OpenStack Summit earlier this year, we released OpenStack Everywhere based on Rackspace Private Cloud, along with an optimized hardware stack and data center colocation services as a managed service. During our initial design phase, we wanted to provide customers with an easier route to multi-cloud, which is why we partnered with Equinix, the world’s largest IBX data center and colocation provider.
Equinix has a 99.9999 percent data center uptime SLA, making it a perfect home for OpenStack Everywhere. Equinix offers Cloud Exchange, allowing for connectivity into all the major cloud service providers.
All this means that if an organization had the desire to leverage AWS public cloud for EC2 compute services, but wanted to retain all their proprietary data in a secure Rackspace private cloud deployment, this could easily be accomplished by leveraging a dedicated path with Cloud Exchange. The use-cases are unlimited.
Learn more about how Rackspace and Equinix’ partnership can benefit your business by tuning into our free webinar, “Is Your Cloud Strategy All Grown Up? Cloud Maturity Through Hybrid-Cloud-as-a-Service” on Oct. 6 at 11 a.m. Central, where experts from both companies will describe how to build a truly mature multi-cloud solution.
What problems can OpenStack Everywhere along with Cloud Exchange solve?
If an organization is required to be located near its customer base, this can be accomplished with OpenStack Everywhere, given the nearly unlimited location flexibly thanks to Equinix’s global data center footprint. At the same time, by leveraging Cloud Exchange, an organization would have a dedicated route into a variety of cloud service provieders. In the use-case mentioned earlier, the scenario of AWS providing additional on-demand compute resource while working in tandem with Rackspace Private Cloud powered by OpenStack, the private cloud would effectively provide storage (block or object), and local data services.
As for our analytics engine, which is “burstable” by nature, one can leverage AWS and EC2 compute. The algorithms that comprise most modern analytics platforms generally do not operate within a persistent state. In this case, AWS EC2 reserved instances could be an ideal fit given the fixed cost and predictability of the analytical running state. Additionally, the AWS analytical workloads could be logically isolated and only made available to Cloud Exchange by using an Amazon Virtual Private Cloud. The technical use-cases are nearly unlimited, especially when pairing Rackspace Private Cloud with AWS for a multi-cloud approach.
But perhaps the most valuable aspect of OpenStack Everywhere is the financial flexibility it can bring to organizations struggling to untether their existing capex spend in the data center. Because OpenStack Everywhere is offered as a managed service, an organization could consume private cloud services, with an optimized hardware stack and all the datacenter economics as an opex model.
To learn more, please attend our Oct. 6 webinar on building the right multi-cloud strategy for your organization.