While hyperconverged infrastructures (HCI) offer rapid startup they lock organizations into a specific set of vendor compute, storage, networking and hypervisor configurations. And, as the infrastructure scales it becomes increasingly difficult to deliver workload specific levels of performance and data protection.
Webinar: Hyperconvergence is Broken, Learn How to Fix it!
1. In this webinar you will learn:
• What are converged and hyperconverged infrastructures?
• How these infrastructures limit flexibility and scalability
• How these infrastructures don't address operational challenges
• What needs to change so they can solve enterprise problems
Hyperconvergence is Broken!
Learn How to Fix it
Webinar
For audio playback and Q&A go to:
http://bit.ly/HCIBroke
2. Our Speakers
George Crump is the founder of Storage Switzerland, the leading storage analyst
focused on the subjects of big data, solid state storage, virtualization, cloud computing
and data protection. He is widely recognized for his articles, white papers, and videos
on such current approaches as all-flash arrays, deduplication, SSDs, software-defined
storage, backup appliances, and storage networking. He has over 25 years of
experience designing storage solutions for data centers across the US.
Tushar Agrawal is in Product Management at Datrium. He joined the team in 2016.
Prior to Datrium he worked in Hyperconverged Product Management at Atlantis
Computing and spent 5 years at HP Servers in HP Integrity & ProLiant Product
Management.
3. • Bare Metal
• 1 App to 1 Server
• Initial Virtualization
• software added to servers
• Convergence
• pre-integrated servers
• Hyperconvergence
• overly integrated servers
• Open-Convergence
• correctly integrated servers
The Evolution of Virtualization
4. • Existing Technology
• pre-integrated
• Helped with time to first VM
• Carried forward operational
challenges
Converged Infrastructure
Compute
Network
Storage
5. • Helped time to first VM
• Overcompensates for
legacy operational issues
• Creates new operational
issues...
Hyper-Converged Infrastructure
Compute, Network, Storage
6. • Increase in internode traffic
• data protection
• communication
• Interdependence between nodes
• Can’t scale capacity independent of performance
• Node down = storage performance or
protection degradation
Hyper-Converged Operational Issues
7. • Despite the “whiteboard” appeal of
Hyper-converged, dedicated storage
still dominant
• Hyper-converged is the right idea but
wrong execution
• Fix the execution with Open-
Converged
The Result
8. • Architecture
• Compute node with compute,
network, flash storage
• Data node with NVRAM, hard
disk
What Is Open-Converged?
9. • Fully Optimized Flash
• Active data written and read from local flash
• Distributed Services Execution
• Compression, deduplication, encryption,
erasure coding, snapshots computed locally
Two Layer IO
10. • Data Shared and Protected
• All writes to shared data node
• All data always on shared data
node
• Persistent copy of data stored in
data node and globally
deduplicated
Two Layer IO
11. • Compute nodes no longer reliant on each
• Nodes become stateless again
• Nodes can be different types
• Independently scale compute & capacity
• Software-defined, server-powered data services
(eliminate central bottlenecks)
Why Do We Call It Open?
16. Datrium DVX Rackscale
Open and Converged
Effortless Private Cloud Infrastructure
• Turnkey Compute and Data Nodes
• Open to 3rd Party Servers: No Silos
• Minimum: 1 Compute Node, 1 Data Node
Effortless Data Management
• Granular, policy-driven w/ dynamic binding
• Multi-site, WAN-optimized replication
• End-to-end encryption with data reduction
DVX Compute Nodes
For Speed
• Local flash reads
• CPU for data services
DVX Data Nodes
for Persistence, HA
3rd party servers
with DVX Software