Happy summer! It’s hard to believe that six months have passed since our first CrowdChat on #LocalData. Since we’re just as passionate about hybrid cloud as ever, we revisited old questions and asked new ones for last week’s CrowdChat.
Our industry experts agree that adoption of hybrid cloud has accelerated since last December, and aren’t shy about sharing where and how that adoption takes place. Under the umbrella of "Tiering to Public Cloud: How to think about and manage tiering from the private data center to public cloud," we talked about data types, decision factors, and dairy products. You’ll have to read the full transcript to get the last reference!
Here is a summary of our favorite quotes from the founders, CIOs and CTOs that made up the panel on June 22.
Acceleration of Hybrid Cloud
Although Chris Harney of CyberNorth states that the industry still has a ways to go, and that “not everyone has taken the time to get comfortable with cloud or which cloud platform to choose,” the most resounding consensus from the panel came when John Furrier of SiliconANGLE asked for a read on whether hybrid cloud strategies have accelerated or decelerated over the last six months.
As consumers become accustomed to the “multicloud” model, in which more than one cloud vendor is used, and the hybrid cloud model, in which at least some data is on premises, both models must retain some flexibility, maintains Stu Miniman of SiliconANGLE.
Igneous’ own Steve Pao points out that hybrid cloud is currently gaining interest from two camps: enterprises that have established on-premises architectures and “those that think of themselves as ‘cloud first’ ”. Andrew Miller of Rubrik and Chris Dagdigian of BioTeam both cite the former camp’s enthusiasm for cloud technology as a big driver for hybrid cloud. In order to see value from cloud investments, enterprises must integrate them with existing on-premises applications, and do their due diligence ahead of time to ensure that the basic math adds up.
Hybrid Cloud Use Cases
John Furrier asks what applications are driving hybrid cloud use cases.
Pao, Dagdigian, and Miller cite backup as a logical use of hybrid cloud, as on-premises secondary copies of data with tiering to cloud combines on-premises and cloud so smoothly. Miller points out that archive is also an established use case, as many companies end up using cloud archives in place of tape. Keith Townsend of AbbVie agrees that backup and archive are the standard use cases for hybrid cloud, at least for now.
Although the panel reached a consensus that cloud bursting is a potentially very valuable use of hybrid cloud, panelists do not all agree on its current degree of usability. Jeff DiNisco of P1 Technologies shares a success story of a visual effects studio that “burst to cloud for everything” and kept costs down, while Miller remains skeptical, saying that cloud bursting is “great in theory but still really hard in practice given the weight of data and LAN vs. WAN speeds.”
Dagdigian also mentions “federated ID management across multi-organizations” as a top driver of hybrid cloud use.
Keith Townsend of Pluralsight and Furrier mention the shift to consumption of infrastructure as-a-Service. Although it is harder to demonstrate the benefits of IaaS than, say, SaaS, the integration of an application layer into infrastructure makes this more obvious to customers.
“All of the serious providers are offering colocation of the ecosystem of enterprise apps,” Townsend states.
In fact, the panel had no trouble naming about a dozen.
All Types of Data are Relevant
Server backups, SQL databases, NoSQL databases, and unstructured data are all involved in hybrid cloud and cloud tiering.
Of course, as we talk about tiering data to public cloud, the type of data remains relevant. Although all are important to business functions (to varying degrees, based on the enterprise, of course), Pao, Dagdigian, and DiNisco point to file system data as particularly cumbersome.
Since “unstructured data is just a bigger [percentage of the total] data...79% according to IDC and validated by customers,” Pao says, the problem of finding a backup and archive solution that is more reliable than tape over the long term shouldn’t be underestimated. Miller brings up the temptation to use tape backups merely for compliance purposes, with no real confidence that files can be restored to primary tiers or transferred over to new LTO formats, and Furrier agrees that this confidence gap spurs many to look to cloud/cloudlike alternatives.
DiNisco recommends services like Amazon Web Services’ Relational Database Service for structured data, while Steve Pao is interested in Datos IO and its backup/cloud-tiering app for NoSQL databases.
Harney predicts that it will be a long time before local data disappears, and introduces a discussion of vendor lock-in. Pao and DiNisco agree, with Jeff arguing that “a customer loses all leverage once they get to multiple PBs in a single cloud.” PluralSight's Keith Townsend echoes this argument, saying that since “even born-in-the-cloud companies moved back to all-private cloud, ... public vs. private should be revisited as business dictates.”
Essentially, although the total cost of ownership of a small to medium datacenter might be similar to that of using public cloud, Townsend suggests that the loss of control and hassle of moving back on-premises as the datacenter scales to petabytes merits careful consideration. As Dagdigian reminds us, storing a petabyte on premises is significantly less expensive than storing it in public cloud.
Implementation of Hybrid Cloud
Furrier asked the panel to describe the technical implementation of the hybrid clouds they are seeing: What components are involved, and has this changed in the past six months?
Rubrik's Miller points to three main components:
- some level of automation/orchestration
- focus on common data movement/management tools between on/off-premises
- tools for visibility into all environments
Townsend and Dagdigian agree, though Dagdigian would prioritize identity management and Townsend emphasized that “there’s no single pane of glass for hybrid cloud” and therefore a focus on “monitoring and ops technologies that integrate” should be prioritized. Pao agrees that identity management for user-facing apps is especially difficult to implement, and posited that this is responsible for the role of IT-facing applications, such as backup, in early hybrid cloud adoption.
Underscoring the tight relationship between hybrid cloud and multi-cloud, panelists are quick to name specific public cloud providers’ individual strengths, and mention the willingness of enterprises to pick the best tool for the job — whether it’s one public cloud provider, another, or an on-premises solution. Mark Thiele of Apcera says that “latency will continue to be a major driver to choices for workload location and provider”, and that if APIs stay relatively consistent across platforms, management overhead can be minimized.
As Furrier said, “Igneous is on the cutting edge of a huge architectural shift in enterprise computing.” For further reading, download the new Taneja Group whitepaper to learn more about how Igneous solves the problem of file systems backup at scale and get a sneak peek into one of our customer stories. Visit the Igneous Backup page for details on our fast, easy backup and restore solution for unstructured data.
Check out the blog posts "Are Data Backups Broken?", "Managing Billions of Files", "Spawning a New Hybrid Cloud", and "When Data Can’t Move Offsite" to read our takes on previous Igneous CrowdChat conversations. Hope you can join us at the next one!