SwiftStack

Some interesting phenomena occurred this year in the data center. Greater consumption of storage is never a surprise, though it is surprising to see that revenues for traditional storage hardware manufacturers are down. How is this possible? One explanation comes from Amazon Web Services (AWS), who first reported very strong growth in its Simple Storage Service (S3) offering this year, which already earns greater than $1 billion on an annualized basis. Large cloud providers have huge buying power and pay less for commodity materials such ashard drives, and they do not buy marked up storage systems. 

“Public cloud services are similar, as the value of cloud is the standard ability to mix-and-match resource use on-demand.” 

Moreover, enterprises are moving some data center storage workloads to public clouds at a lower cost that would have otherwise gone on classic storage at a higher cost. This is evidence that cloud storage is good enough for some workloads, but public cloud storage is not the only disruptor. If an offering is easy to consume, let’s users pay as they grow, and scales very large, it just has to be good enough for enterprises to start using it. 

Open Source Already Works Wonders in the Data Center 

Once dismissed by proprietary competitors as “immature,” open source technologies from operating systems to middleware, application frameworks to databases, are the standards in enterprise and web infrastructure. This is nothing new. The open source model has completely and fundamentally transformed the infrastructure tier in the data center so that few proprietary infrastructure platform technologies any longer have a sustainable advantage. 

Being “open” is not the biggest reason though. While architects and executives like having flexibility and avoiding lock-in, what they want most from open source in their private infrastructure and for their internal customers are the same characteristics that public cloud provides. They want genuine simplicity in consumption, a model that lets their customers pay for what they use, and an architecture that is multi-tenant and free of technical limits that might prevent future scale. 

An Example - OpenStack 

OpenStack is cloud software that is built on open source. It has been powering large public clouds such as Rackspace, HP, IBM Softlayer and others for years. Using open source platforms is an irreversible change in enterprise IT architectures. Some projects in OpenStack are relatively new, while others are tested and proven at scale. This multi-tenant scale is often an order of magnitude larger and more demanding than an individual enterprise deployment. 

In the context of the AWS S3 example, the OpenStack Swift project is a technology that service providers and enterprises are also using as part of their own private storage clouds. Applications and end users get the same “cloud” experience, while IT owns and manages data and services from within their data centers, behind their firewalls and private networks. Not only is it good enough, it is working wonders for organizations of all sizes. 

The big disruptor here is standard hardware. This helps explain how consumption of a resource like storage can be at an all time high, while revenues from storage hardware companies is down. Most enterprises do not have the buying power of a large cloud operator, but they can procure commodity hardware at market prices. Unlike traditional storage systems, the market price for commodities such as standard x86 servers and standard hard disk drives has little inflation built-in. 

What’s more, enterprises are already very efficient at buying server infrastructure. They often have a volume purchase agreement with their favorite vendor, or they routinely purchase standard hardware from multiple vendors as part ofa strategy that lowers cost and assures cost certainty when scaling. Public cloud services are similar, as the value of cloud is the standard ability to mix-and-match resource use on-demand. On the hardware side, enterprises are already doing mix-and-match hardware in their private data centers. 

The Same Example - The OpenStack Marketplace 

OpenStack software runs on standard server hardware and standard hard disk drives. Using open source storage software does not preclude organizations from working with commercial vendors. Organizations using big data are turning to commercial vendors such as Cloudera, HortonWorks, and Mapr, but also can turn to technology vendors they already work with such as EMC, HP, and IBM. The OpenStack ecosystem has a similar if not more robust set of participants. These vendors do not position their OpenStack offerings as “immature.” 

Later this month the OpenStack community will gather in Tokyo for its next semi-annual global summit, number 12 in the series. Vendors in the ecosystem will exhibit their wares in the marketplace. Businesses will continue serving users in the way that users prefer to consume those services. Vendors are offering more solutions to their customers with the advantages of cloud, as the challenge is not just to do more, but alsoto do more with the same or even fewer resources. 

In conclusion, public cloud delivers what enterprises need in a way that is good enough. Those enterprises must maintain custody and control of many applications and data, and do it in a way that delivers the cloud experience. As with any disruption in technology, new models will continue to develop. Models that deliver on what users want and run on what enterprises already know how to consume are bound to have continued success.