Upload
others
View
13
Download
0
Embed Size (px)
Citation preview
2
Agenda
• Enterprise Storage Business Challenges
• SUSE Enterprise Storage Architecture
• SUSE Enterprise Storage 6
• SUSE Enterprise Storage
Use Cases
• Questions
33
Enterprise Storage Business Challenges
4
The Data Explosion Continues
Mobile Data
Emails
Transactional
Data
Videos
Medical Data IoT Data
175 ZBby 2025
5
Limiting Factors of
Traditional Enterprise Storage
Difficult to Scale and
Manage Data Growth
Expensive Won’t Extend to
the Software-defined
Data Center
$
66
SUSE Enterprise Storage Architecture
7
Powered by Ceph TechnologySUSE Enterprise Storage Architecture
88
SUSE Enterprise Storage 6
9
SUSE Enterprise StorageLast 12-month Accomplishments
Monitor
Nodes
Management
Node
Storage
Nodes
Unified
Open Source
Software on x86
and Arm
Resilient &
Self-healing
High
Performance
Massively
Scalable
Public Cloud
Like Pricing
Object
Storage
Block
Storage
File
System
Unified
Cluster
Hardware
Flexibility
Reduced
IT Costs
• Launched SUSE Enterprise Storage 5.5
o 8th release
• Latest upstream release Ceph Nautilus
o 8 out of top 20 Ceph contributors
are from SUSE
• SUSE team driving Ceph Dashboard
upstream project
• SUSE more than doubled Ceph
customers in FY18 relative to FY17
10
SUSE Enterprise StorageDevelopment Focus Areas
• Unified Block, File
& Object
• Fabric
Interconnects
• Cache Tiering
• Containerization
• Hierarchical
Storage
Management
• Backup/Archive
• Continuous Data
Protection
• Remote
Replication
• Ease of Installation
• GUI based
Monitoring &
Management
Manageability Interoperability Efficiency Availability
11
SUSE Enterprise Storage 6
** Items are tech preview
• Ceph Dashboard (oA replacement)
• Phone home (metrics and error
analysis)
• Enhanced PG balancing
• Autonomous PG scaling**
• CephFS directory quotas
• Graceful system shutdown
• Storage enclosure disk integration
Manageability
12
• IPv6
• RGW Elasticsearch sync module
• Node to node encryption**
• Multiple CephFS instances in one
cluster**
• NFS-Ganesha active/active
architecture**
Interoperability
** Items are tech preview
SUSE Enterprise Storage 6
13
• QoS for RBD
• QoS for background operations
• S3 object policy driven tiering**
Efficiency
** Items are tech preview
SUSE Enterprise Storage 6
14
• Unidirectional sync
external cloud via S3
• CephFS snapshots
• Asynchronous file replication**
Availability
** Items are tech preview
SUSE Enterprise Storage 6
15
SUSE Enterprise Storage Roadmap
*: These dates are indicative only, not guaranteed
Unified File, Object & Block
• CephFS
• Multisite object replication
• Acquired openATTIC
Performance and Ease of Use
• BlueStore• Compression• Block/File
Erasure code• NFS/S3 interface• NFS/CephFS interface• OpenATTIC integrated
with DeepSea• Enhanced monitoring
and management
Interoperability and Security
• Integration with SUSE OpenStack Cloud
• CIFS/Samba interface• Non SUSE
Ceph clients• Event notification• AppArmor integration• Security audit testing
Enhanced Management and Hybrid Cloud
• Ceph Dashboard
• Enhanced monitoring
• Public cloud product
• Sync to external cloud
• Phone home
• RBD QoS
• CephFS snapshot
• Async file replication
Containerization and Autonomous Self-Healing
• Containerization
• Enhanced intelligent data placement
• Guided troubleshooting
• Guided HW upgrade
• Guided OSD repair
• Data deduplication
• RDMA interconnect
• Windows drivers
SES 4: Nov 2016 SES 5: Oct 2017 SES 5.5: Oct 2018 SES 6: June 2019* SES 7: Q2/Q3 2020*
16
2017 2018 2020 2021
SUSE Enterprise Storagev4
v5
v6
Built On • Ceph Luminous release
• SUSE Linux Enterprise Server 12 SP3
Manageability• openATTIC phase 2
Grafana monitoring dashboard
Prometheus event alert - email
• DeepSea (Salt) phase 2
Online Filestore to BlueStore
Interoperability• NFS Ganesha
• NFS access to S3 buckets
• CIFS Samba**
• CephFS Multi MDS support
Availability• Erasure coded block and file
Efficiency• BlueStore back-end
• Data compression
v5
Built On • Ceph Luminus release
• SUSE Linux Enterprise Server 12 SP3
Manageability• Internationalization
• Usability enhancements
• Predefined profiles
• Autonomous data balancer**
• DeepSea (Salt) phase 3
Interoperability• Non SUSE RBD and CephFS clients
• CIFS/Samba
• AppArmor security module
Availability• Multisite RADOSGW N+1 with N/N-1
Efficiency• BlueStore/RocksDB optimizations
Built On • Ceph Octopus release
• SUSE CaaS Platform
Manageability• Integration with Rook + Kubernetes
• Phone home phase 2
• Autonomous PG scaling
• Self-healing PG
Interoperability• Containerized deployment
• Native Windows client drivers
• Node to node encryption
• Multiple CephFS instances in one cluster
• NFS-Ganesha active/active architecture
• RDMA back-end**
Availability• Bidirectional sync external cloud via S3
• Asynchronous file replication
Efficiency• Data deduplication
• SW cache layer
• Object Storage Daemon optimizations
• S3 object policy driven tiering
Built On • Ceph Nautilus release
• SUSE Linux Enterprise Server 15 SP1
Manageability• Ceph Dashboard (oA replacement)
• Phone home (metrics and error analysis)
• Enhanced PG balancing
• Autonomous PG scaling**
• CephFS directory quotas
• Graceful system shutdown
• Storage enclosure disk integration
Interoperability• IPv6
• RGW Elasticsearch sync module
• Node to node encryption**
• Multiple CephFS instances in one cluster**
• NFS-Ganesha active/active architecture**
Availability• Unidirectional sync external cloud via S3
• CephFS snapshots
• Asynchronous file replication**
Efficiency• QoS for RBD and background operations
• S3 object policy driven tiering**
V5.5 V6 V7
* Information is forward looking and subject to change at any time.
2019
v7
** Items are tech preview
17
SUSE Enterprise Storage — Management
Install• Import unmanaged SUSE Enterprise
Storage cluster
Configure• Online Filestore to BlueStore
Manage• Deploy, configure, assign nodes
• iSCSI target management
• RGW users, buckets, keys
• View, create, delete CephFS
• NFS Ganesha gateway
Monitor• Grafana dashboards
• CephFS status
Diagnose• OSD health status
• Configure• Convert EC to replication (visa versa)
• Manage• RBD snapshot
• OpenStack integration
• Cluster rebuild progress
• Admin node redundancy phase 1
• Monitor• Event notification alerts**
Configure• Deploy new OSD to existing node
• CIFS/Samba
• RGW and CephFS remote replication
• Sync to external cloud
Manage• CephFS snapshot
• Cluster topology map
• Guided upgrade process
Monitor• OSD and MDS statistics
• Ceph client performance
Diagnose• Predictive disk failure analysis
• Guided troubleshooting
• Guided OSD repair process
Configure• iSCSI
• NFS-Ganesha
• Phone home
• RBD QoS parameters
• Background operation parameter (QoS)
• RBD remote replication
Manage• Role-based Ceph Dashboard
• Localized Ceph Dashboard
• Ceph Dashboard SSO (SAMLv2)
• Crush map viewer
• RBD snapshot
Monitor• Real-time performance metrics
• Scale-out Prometheus monitoring
• Event notification alerts
• SNMP traps
• Command audit log
2017 2018 2020 2021v4
v5
v6
v5 V5.5 V6 V7
2019
v7
* Information is forward looking and subject to change at any time.
** Items are tech preview
1818
SUSE Enterprise Storage Use Cases
19
Backup to Disk Solution
Custom Build
DIY with COTS
Servers
Appliance
HPC Storage
Certified Reference Architectures
Cloud & App
Delivery
SUSE OpenStack
Cloud
Container as a
Service
Analytics Custom Apps
SUSE Enterprise Storage
Cla
ss
ic
Wo
rklo
ad
s
Clo
ud
Na
tive
Wo
rklo
ad
s
Data Protector
Use Case Focused SolutionsPartnership Ecosystem
Fle
xib
le
Co
nfi
gs
File Sync and
Share
2020
Private Cloud: Mixing Evolution with Revolution
2121
Do I Even Need to Change?
22
23
If You’re Not Innovating, You’re Falling Behind
Keep doing what
you’ve been doing…
Competitive
innovation
24
Competitive
innovation
More like this!
Keep doing what
you’ve been doing…
If You’re Not Innovating, You’re Falling Behind
25
Do I Even Need to Change?
End-users are ALREADY Changing…
• How many “shadow IT” workloads are already running
on AWS/Azure/Google/etc.?
Culture vs. Doctrine
• Are we doing this because it’s the path of least
resistance, or are we doing this because it’s right?
• If budget and politics are removed, would we still do
things this way?
IT Marketplace is Changing
• If everyone else is changing, am I prepared for my next
job (internal or external)?
• Would my team rather focus on improving repetitive,
routine tasks or on enabling creativity?
2626
Why Haven’t We Done This Already?
27
Fear is the Mind Killer
“[Fear] will only
make someone
work just hard
enough not to
get fired.”
– Peter Gibbons (Office Space)
28
Why Haven’t We Done this Already?
Can’t abandon what’s making
you money as you transition
to the “next big thing.”
29
Why Haven’t We Done this Already?The Modern Enterprise Data Center
Mixture of Archaeology and Technology
• Finance has Cobol from the 60s still in production
• Aerospace has product lifecycles measured in decades
• Platforms emerge over years, rarely go away
• zVM
• VMware
• Xen
• KVM
• LXC
• Docker
• CRI-O
30
Why Haven’t We Done this Already?Existing platforms are more than just their software
• Home-grown scripts for
managing it.
• Regulatory compliance
• In-house workflows and
policies
• IT administrators have skill
sets targeted to existing
infrastructure
What is the
greatest barrier to
IT transformation?
32
3333
How Can We Fix This?
34
The Bridge You Need
35
OpenStack does Mode 1
Script/CLI
Dashboard
OpenStack
API Endpoints
Platform
• Bare Metal
• Virtualization
• Containers
Network
• Dynamic custom IP space
• Load Balancing
• VPN
• Firewall
• DNS
Storage
• Block
• Object
• Shared
Orchestration
Optional
User
36
OpenStack does Mode 1
Without letting users run wild
• Secure, isolated
multi-tenancy
• Quota restrictions
37
OpenStack enables Mode 2
IT operations can pre-
emptively define the
limits for cloud-native
application development
Developers focus on
application development,
not plumbing
38
OpenStack
Frees IT
38
• Reduces risk of
choosing platforms
• Simplifies migrations
• Allows IT to be more
aggressive in
adopting new
technologies
39
OpenStack Frees IT
All with NO
• Rewrites of scripts using
OpenStack APIs
• Change to existing
workflows or policies
• Change to end-user experience
• Need to retrain everyone
40
For example…
Catalog
User
4141
Would you like to see a demo?
42
Every single
demo you’ve
seen today
used it.
4343
Enter the chameleon…
44
Enter the chameleon…
• Built on SUSE Linux Enterprise
• Expertise to get you running, and
support to keep you running
• Established partnerships to expand
the functionality of your private cloud
• Stable platform with the flexibility
necessary for modern applications
SUSE OpenStack Cloud
45
SUSE Provides Flexibility in PLATFORM
• KVM
• VMware
• z/VM
• Containers
• Bare Metal
46
SUSE Provides Flexibility in NETWORKING
• Linux Bridge
• Open vSwitch
• VMware DVS
• VMware NSX
• Cisco Nexus
• Cisco ACI
47
SUSE Provides Flexibility in STORAGE
• Raw/Local
• Ceph
• NetApp
• EMC
• EqualLogic
• Fujitsu
• Hitachi
• Pure
• VMware
48
SUSE Enables Installation
a
• Installation measured in minutes, not months
• Shortest POC setup = 6 hours (includes unboxing, rack &
stack, cabling)
• Typical length of POC is three days
• Integrated HA cluster deployment for control plane and KVM
instances
“We have more done in three days than we had
in the last six months of doing it ourselves.”
49
SUSE Enables Upgrades
First to offer fully-supported, in-place upgrades
• Since SUSE OpenStack Cloud 2 (Grizzly)
As of SUSE OpenStack Cloud 6 (Liberty), we
support zero-downtime upgrades
50
“We don’t just drop
it off at the end of
the driveway…”
5151
Technical Overview & Roadmap
52
What’s New in SUSE OpenStack Cloud 9?
• Based on OpenStack Rocky• Multi-attached storage
• Ironic Improvements
• Includes Queens functionality
• SUSE Linux Enterprise Server 12 SP4
• Day two UI – CLM Admin Console
• IPV6 Support**
• Watcher Optimization – Tech Preview**
• GA release planned for April 2019
• Dual lifecycle options• Crowbar
• Cloud Lifecycle Manager (CLM)
**Delivered in Cloud 9 Updates
53
Day 2 UI SUSE OpenStack Cloud 9CLM Admin Console
54
What is it?• OpenStack services (e.g., Nova, Keystone, Neutron) packaged to run in containers as
opposed to running as processes on bare metal
• Kubernetes installed onto bare metal, with OpenStack deployed within containers using
Helm Charts
• Airship open source project foundation for lifecycle management
Why are we introducing it?• Containers are more lightweight than VMs and easier to start up and terminate
• OpenStack environment faster to start up and scale
• Easier to separate individual OpenStack projects
• Easier to scale individual components up and down
• Upgrades of individual components become easier
• Leverage Kubernetes’ built-in HA
• Designing a self-healing environment becomes simpler
• Securing individual services and components becomes easier, due to container isolation
Containerized OpenStack Tech Preview
55
2018 2019 2020 2021
SUSE OpenStack Cloud
8
9
10
8
Built On• OpenStack Pike Release
• SUSE Linux Enterprise Server 12 SP3
New or Expanded Services• SUSE CAP Integration
• Physical Server as a Service (Ironic)
• SDN Support for NSX-V
• Dual lifecycle manager options
Operational Enhancements• Non-disruptive Upgrade to Cloud 8
• Planning and Pre-install Validation
• Simple Deployment UI
• Scale Testing 200 nodes
• Monitor Capacity and Performance
• 3-year support
Built On• OpenStack Rocky Release
• SUSE Linux Enterprise
Server 12 SP4
New or Expanded Services• Containerized OpenStack**
• Dual lifecycle manager options
Built on• OpenStack Pike Release
• SUSE Linux Enterprise Server 12
SP3
New or Expanded Services• CLM Manila Support
Operational Enhancements • Lifecycle Tools Improvements
• mkcloud support
• SES Integration
8 Updates 9 10
* Information is forward looking and subject to change at any time.
** Items are tech preview
Built On• OpenStack Train Release
• SUSE Linux Enterprise Server 15
New or Expanded Services• Containerized Deployment
• Accelerator Engine Support
Operational Enhancements• Prometheus Monitoring
• Multiple Site Enhancements
• DR Enhancements
• Root Cause
Detection/Analysis/Repair
• Kubernetes Networking
Configurations
• Workflow Automation
9 Updates
Built On• OpenStack Rocky Release
• SUSE Linux Enterprise
Server 12 SP4
New or Expanded Services• GPU Support
• SDN Support for Juniper
• SDN Support for NSX-T
Operational Enhancements• IPV6 Support
• Policy-based Optimization**
• Scalability Improvements
-Region Support
-Federation
-Multi-Data Center Support
• Cloud Monitoring
-Lifecycle Events Monitoring
-Advanced Log Analysis
-Monitoring Analytics
• Integration with SUSE Single
Sign-on
5656
Summary
57
OpenStack lets you define
the game, then gives users
the freedom to play.