Storage 2015 Storage Shifts and Software Defined Storage (SDS) MRMUG Chris Walker Solution Architect cwalker@ciber.com Ciber
Storage 2015 BIO Ciber – Solution Architect • 27 years working in IT in an assortment of roles including: Systems Administrator, Programming, Software Engineering, Critical Problem Support, technical Sales, and most currently a Solution Architect for Ciber, Inc. • Tennis, Guitar, Jet Skiing(new), going out to eat! • I have lived in many different places but most recently have lived in the Boston area and moved to Northern Kentucky a little over 3 years back. I am married and have one daughter who recently graduated college and I just bought the farm!
Storage 2015 Agenda – Presentation Topic Outline • The New Storage Directions and Driving Forces • Software Defined Storage – What is it? • How can it be done? • Why do it? • An Example of SDS from the IBM portfolio • Converged and Hyper-converged IT • Q&A
Storage 2015 4 3 2 1 5 0 25 32 31 30 29 28 27 26 60 24 23 19 18 17 15 14 10 33 34 35 36 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 44 43 42 41 40 39 38 37 11 9 8 21 12 13 16 20 22 6 7 Gone In 60 Seconds 347,000 Tweets 693,800 GB of IP Data is Transferred 571 New Websites Created Over 2 million Google search queries 47,000 Apple Store App downloads 77,244 Wi-Fi connections made on iPhone 217 New Web Mobile Users 600 New Videos on YouTube 60 www Domains Registered 20 New Victims of Identity Theft 1,500 Blog Posts More Than 200 Million E-Mails 510,000 Comments on Facebook More Than 370,000 Minutes of Voice Calls on Skype
Storage 2015 Welcome to the world of…. C loud, A nalytics, M obile and S ocial Mobile Cloud Data 95% of mobile 80% of new applications Explosion traffic is data will include cloud delivery 90% or deployment 2.5 Billion of data Gigabytes of created in Social data per day last two years 500 million Tweets a day; 7 million apps and websites integrated with Facebook Data doubling every two years Big Data & Analytics Today ’ s Business CAMS 2,500 petabytes of big data are being generated every day 4
Storage 2015 Systems of Interaction The new era of transforms how companies engage, interact and transact with their customers. Systems of Interaction 20% Mobile/Social 80% of new data of new data Traditional IT Exploding Diverse Increasing Value Mobile and Social Data Volumes Data Types of Information Engagement (Big Data) (Big Data) (Analytics) Block Data = Structured Data (Data Bases, VMs, core business applications) File and Object Data = Unstructured Data (Text, Video, Pictures, Office Docs, Device data, Social and Mobile conversations)
Storage 2015 Systems of Interaction in Play Outcome : The Right Car in the Right Place and the Right Time for the Right person Sales Campaign launch – new C-Class Sales Campaign launch – new C-Class Conversations and Social interactions with friends and family around event. Comments on research, word of mouth, etc, all about Mercedes and even their competition. Tweak the Adjust Campaign Inventories, Search information driven from Sales campaign Logistics -Reviews -Colors Big Data -Options Business Intelligence (Analytics) Locations known for audience! Traditional IT
Storage 2015 Behind the Scenes is IT supporting this… If you were to choose .. Which is most important in this list? Lines of Business Driven LOB accounts for 61% of all IT purchases CMO drives a majority of Mobile/Social/Analytics most often without IT involvement
Storage 2015 Servers are well understood. Storage is exploding Pressure to adapt to a fast and changing business environment Hybrid Cloud / Public Business New Data Cloud Pressures Applications Growth IT Need a new storage model that : • Frees data from constraints of specialized hardware • Lowers overall Costs • Provides location independent data services • Faster deployment • Ease of management and growth • Integration with modern APIs and hypervisors • Enables Faster time to market!
Storage 2015 What is Software Defined Storage (SDS)? Software-defined storage is any storage software stack that can be installed on any commodity (x86 hardware, hypervisors, or cloud) and/or off-the-shelf computing hardware and used to offer a full suite of storage services and federation between the underlying persistent data placement resources to enable data mobility of its tenants between these resources Key attributes of SDS Runs on commodity hardware: no special hardware or components needed Offers a full suite of storage services: equivalent to traditional hardware specific systems Federates multiple persistent storage resources: internal disk, cloud, other external storage System Leverages: – Shared nothing architecture – Scale-out model – Virtualization of Storage resources Broadly grouped into three buckets: File-based, Object-based and Block-based
Storage 2015 Software Defined Storage - Basics – Control Plane and Data Plane Control Plane (Software) Network of Servers with Storage (Back-End Could be anywhere running on almost anything) Software Defined Storage Traditional Appliance Data Plane “Integrated device” (Hardware)
Storage 2015 Software Defined Storage – Control and Data Planes Concepts SDS Control Plane (SOFTWARE) Storage Backup and Integration and Storage Analytics and Policy Efficiency Copy API Services Optimization Virtualization Automation Managemen t Open Common Control Plane to Manage Storage Services for application Workloads Monitor Charge Service Self- Provision Life Cycle Security Classes back Service and Storage Interface Manage (GUI) SDS Data Plane (HARDWARE) Scalable Storage Performance and Capacity Virtual Storage Storage Systems
Storage 2015 Software Defined Storage – Control and Data Planes How vendors deploy today - Typically SDS Control Plane (SOFTWARE) Backup and Integration and Storage Analytics and Storage Policy Copy API Services Virtualization Optimization Efficiency Automation Managemen t Open Common Control Plane to Manage Storage Services for application Workloads Service Charge Self- Security Provision Life Cycle Classes Service back Tight Coupling SDS Data Plane (HARDWARE) Scalable Storage Performance and Capacity How we deploy Storage IBM, EMC, NetApp, Hitachi, HP today Systems Specialized, fit to task Hardware Devices “Storage Arrays” Unified Block File, Object
Storage 2015 Software Define Storage – Control and Data Planes How the industry is starting to deploy and future direction SDS Control Plane (SOFTWARE) Backup and Integration and Storage Analytics and Storage Policy Copy API Services Virtualization Optimization Efficiency Automation Services Open Common Control Plane to Manage Storage Services for the Workloads Service Charge Self- Provision Life Cycle Security Classes Service back Decoupling of HW and SW SDS Data Plane (HARDWARE) Scalable Storage Performance and Capacity Customer picks their Deployment Model: Tape Libraries, SSDs, Flash HW Integrated Appliance (SW+HW), SaaS Cloud, Software Only appliances, Disks to deploy on their own hardware platform File, Block, Object
Storage 2015 Software Defined Storage – How Does it Work? • Most Storage arrays are: • Built on Linux, Unix or Windows Kernels - potentially offering a common code base • Most Run on Intel x86 – offering a common hardware architecture • All are based on Server technology at the core – While delivering advanced data services, most storage products are essentially based on 2 or more servers under the covers. • What are the components of a server? • CPU and associated architecture • Memory • Disk Bays (SSD, Hard Drives) • PCI BUS- I/O • HBAs – Fibre Channel, NICs, SAS • Other PCIe cards as needed IBM V7000 Gen 2 Controller node
Storage 2015 Software Defined Storage – How Does it work? • OK, so storage services are essentially delivered from a server platform. How does one certify all the possible server platforms out there?? • Virtualization\Hypervisor – if the server platform is VMware ESX, Linux KVM, Hyper-V certified it can be certified for use • No hardware compatibility lists or expensive testing labs • Virtualization vendors have done the interoperability work already Server 1 • Insure required components in the server – software deployed atop a virtual machine or a virtual appliance should perform Server 2 basic component checks • Does it have the required interfaces? NICs, HBAs • Does it have the min amount of memory, drives, CPU cores, etc Server N+1 • Is it highly available? i.e. redundancy in components and paths • Caution – UPS, Flushing of cache to non-volatile devices • Customer responsible for UPS, consistent power Data Services
Recommend
More recommend
