- KEY TAKEAWAYS
- Well-defined APIs and the right level of abstraction are essential for successful web services.
- The traditional approach to storage results in web services at the wrong level of abstraction, the LUN.
- By operating at the VM and container level, Tintri enables storage services that are easy to consume in cloud environments.
This blog is part two of a three-part series on storage and the Enterprise Cloud.
Storage, however, has not followed suit. Storage industry players, old and new, may have some APIs, but their storage doesn’t fit easily with existing compute and networking building blocks. This is because they haven’t delivered on the second element necessary for successful web services; they don’t operate at the right level of abstraction.
Think about any of the popular hypervisors; all operations are at the level of the VM. Storage based on LUN-level abstractions will never fit seamlessly into an environment built around VMs. The same holds true for containers.
The Tintri Approach
Tintri has always taken a web services approach. By operating at the right level of abstraction—VMs and containers—Tintri storage will “snap in” to your cloud environment. A comprehensive set of REST APIs turn our storage services into Lego blocks.
Using this web services approach, Tintri delivers the agility of public cloud inside your four walls, resulting in an enterprise cloud that meets the needs of existing enterprise applications and cloud-native applications.
A web services architecture allows you to do three things that are impossible otherwise:
- Automation
- Analytics
- Self-service
In my next post, I’ll talk more about how Tintri approaches each of these options.
Tintri Enterprise Cloud is tailored to meet the unique needs of the enterprise data center. I want to discuss how and why Tintri’s architecture uses web services as building blocks similar to public cloud.
What Are Web Services?
Web services are the building blocks that underlay cloud. With a web services approach, infrastructure can be managed in a programmatic way, resulting in services that are both extensible and automatable. There are two success factors for web services:
- Well-defined APIs to facilitate integration and automation
- The right level of abstraction—working in the units of cloud (virtual machines, vDisks, containers, etc.), not the units of legacy infrastructure
Done correctly, individual services “snap together” like Lego blocks. And as with Legos, everything fits together easily to become part of a larger design.
For example, Amazon Web Services (AWS) provides services for compute, storage, and many other functions in the public cloud that all work together seamlessly. Amazon CEO Jeff Bezos first mandated that every Amazon service have an API back in 2002. The success of both Amazon.com and AWS can in large part be traced back to this commitment to web services.
For the Data Center, Storage Services are the Missing Link
If you look at infrastructure in on-premises and service provider data centers, both compute and networking already operate as web services. Compute has functioned like a Lego block for a long time—with hypervisors such as VMware (2001), Hyper-V (2008), and others—not to mention the potential of containers. Networking is becoming a Lego block with products such as NSX from VMware and ACI from Cisco.
