Bill Roth, Ulitzer Editor-at-Large

Bill Roth

Subscribe to Bill Roth: eMailAlertsEmail Alerts
Get Bill Roth via: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn

NexentaStor continues to be adopted by cloud providers and by others, including enterprises building cloud-like systems.  Simple question: why?

1. Economics

You cannot purchase and power a legacy storage system, even if you are a great negotiator and can get 70% or greater discount, for the price at which S3 is sold. This impacts both cloud service providers, who do not want their underlying storage, and enterprises whose charge backs are compared at least implicitly to S3. If you want your developers to stop developing on public cloud services, you’ve got to a) be able to provide them with storage more or less on demand and b) at perhaps no more than 5-10x more than Amazon. The days of your developers accepting a 6 week wait for space on the SAN and a chargeback rate 100x S3 pricing are over.

2.  Random I/O solved by NexentaStor with SSDs

The cloud includes virtualization, and virtualization inherently randomizes the I/O patterns. Random I/O can drop the performance of a legacy, disk-centric array by as much as 10x versus synchronous I/O. By comparison, NexentaStor is the best unified storage on the market at leveraging SSDs for high performance and, when SSDs are used with NexentaStor, you can achieve extremely high IOPS. Partners, such as Boston in EMEA and Pogo Linux in the United States, sell NexentaStor-based solutions that achieve 30,000 to 250,000 or more random IOPS in a 3U form factor starting at $15,000.

3. NexentaStor has rest-APIs and a software business model so you can automate everything to achievestorage on-demand

Once you open a cloud, you can be hit with additional storage demand unpredictably. This applies both to cloud computing and cloud storage services. For example, I recently heard of one of the world’s largest online video game companies that decided to use KT’s cloud for some rendering. The result was hundreds of new virtual compute clusters turned-on, without notice.  It’s that “without notice” that differentiates the cloud, and especially storage for the cloud from traditional approaches. The cloud vendor must be able to quickly provision servers, as needed, to the right use case, whether that is compute services, underlying storage for compute, storage sold as a service, or other services such as CDN, in order to address the unpredictable demand of cloud services.

NexentaStor makes this easy. You can establish a gold master NexentaStor reference configuration and image the NexentaStor build onto systems as needed.  You then can manage NexentaStor completely through the rest, along with other APIs. With 3.1, you also can have your management system be alerted on configuration changes; in conjunction with LDAP access control integration, you can insure that your run book is being implemented correctly. Hardware-based systems can take weeks to receive, power-on, network, and so forth. And that’s before you even dive into their interfaces for configurations.  By comparison, with NexentaStor you can pre-deploy a rack of servers and then decide how to image them on demand.

4. Flexibility : block and file storage, multi-level data protection, and the ability to virtualize existing storage

Let’s say you believe that for your use case and particular cloud design that NFS is the right solution. You like being able to treat each VM as a file and that approach simply works for you. But then, for whatever reason, you decide that iSCSI would be best for the VDI service you are offering from your cloud as well. And you’d like to also use NexentaStor internally to provide Window storage.  Meanwhile, your QA team wants to use NexentaStor for VMware Lab Manager via their fiber channel infrastructure.  With NexentaStor, you can do all of the above from the same solution.

And with our soon to be released Name Space Cluster you can also combine many NexentaStor instances into one NFS name space from the perspective of the clients, further improving manageability and enabling improved performance thanks to the ability to identify and migrate load away from heavily loaded NexentaStor nodes.

5.  Nexenta is proven in the cloud

Nexenta now is well regarded as a solution for storage for the cloud. As a result, we are seeing many of the most interesting deployments occurring in North America, Europe, and parts of Asia and Africa as well. This experience makes us a trusted advisor to many hosting and service providers. You’ll see us announcing partnerships with some of the top cloud-oriented professional services companies in the world as a result of this track record. What this means to end users is that Nexenta can help you walk through the trade-offs of different designs at least at our layer of the stack. Our architectural expertise in our layer of the cloud helps our customers build clouds with designs that scale for years.

That’s at least five reasons that I see. Our primary competitor for large cloud deployments are companies building their own file system. We have one major distinction versus self-built or emerging open source file systems. We never lose the data.  Ever.

Let’s call multi-level data protection a bonus #6 reason we’re doing well.

If you’ve built your own file system and run it at scale, you very likely have lost your customers’ data and you’ve found that while building a file system is simple, maintaining it costs millions of dollars a year.

As always, I welcome your feedback.

Read the original blog entry...

More Stories By Bill Roth

Bill Roth is a Silicon Valley veteran with over 20 years in the industry. He has played numerous product marketing, product management and engineering roles at companies like BEA, Sun, Morgan Stanley, and EBay Enterprise. He was recently named one of the World's 30 Most Influential Cloud Bloggers.