Internet DRAFT - draft-sds-definition-overview
draft-sds-definition-overview
Internet Engineering Task Force Eui-Nam Huh
Internet-Draft Kyung Hee University
CDNI Working Group Ga-Won Lee
Intended status: Informational Kyung Hee University
Expires: Aug 16, 2016 Yunkon Kim
Kyung Hee University
Jintaek Kim
Consortium of Cloud Computing Research
Feb 15, 2016
Software-Defined Storage Definition and Overview
draft-sds-definition-overview-00
Abstract
In accordance with rapid increase of data related to IoT and big
data, techniques to control high capacity data is currently active
and vibrant research field. Enterprises are trying to manage data on
the cloud because of flexibility and capability. However, there are
some limits to handle data intelligently in cloud. The SDS is
considered as a good technique regarding this manner. SDS improves
efficiency, scalability and flexibility in scale-out architecture,
as well as, provides a cost effective solution using the existing
storage resources efficiently.
Status of This Memo
This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet-
Drafts is at http://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress."
This Internet-Draft will expire on August 16, 2016.
Copyright Notice
Copyright (c) 2016 IETF Trust and the persons identified as the
document authors. All rights reserved. This document is subject to
BCP 78 and the IETF Trust's Legal Provisions Relating to IETF
Documents (http://trustee.ietf.org/license-info) in effect on the
date of publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License.
Table of Contents
1. Introduction ----------------------------------------- 2
2. Related Work ----------------------------------------- 2
2.1 Survey of existing technologies for SDS ----- 2
2.2 Standardization activities to SDS ----------- 4
3. SDS ------------------------------------------------- 6
3.1 Definition of SDS --------------------------- 6
3.2 Overview of SDS ----------------------------- 8
4. References ------------------------------------------- 10
1. Introduction
IoT, Big Data and mobile paradigms are leading us to data-centric
computing society. One of the objectives of data-centric computing
paradigm is to realize 'Data-as-a-Service' in cloud storage
perspective. A storage federation is good candidate technology to
realize these paradigms. Software Defined Storage (SDS) is one of
storage and data federation technology.
The SDS is a software-based technology that detaches the storage
management from the physical storage and transforms it into a service.
The SDS technology decreases storage management complexities by
providing an automated and centralized management service to
administer. Furthermore, the SDS improves efficiency, scalability and
flexibility in scale-out architecture, as well as, provides a cost
effective solution using the existing storage resources efficiently.
Therefore, potential of storage and data federation on SDS is
compelling. The purpose of this contribution is to launch a new work
item of storage/data federation by describing the overview,
requirements and capabilities for SDS. This contribution will support
to form a logical storage pool, to manage it by software, and finally
to federate storage/data logically.
Through this contribution, various derivation, extension and
combination of services can be created. Moreover, system
interoperability improvement can give us new demand creation between
different domain industries. Therefore, this contribution can
contribute to invigoration of big data industry and expected to lead
cloud and big data markets.
2. Related Work
2.1 Survey of existing technologies for SDS
2.1.1 HP StoreVirtual VSA
Transform server's internal or direct-attached storage into a fully featured shred storage - array without the cost and complexity
associated with dedicated storage.
2.1.2 ViPR
EMC provides intelligent SDS solutions that help organizations
drastically reduce management overhead through automation across
traditional storage silos and pave the way for rapid deployment of
fully integrated next generation scale-out storage architectures.
2.1.3 IBM Spectrum Scale (GPFS)
IBM Spectrum Scale is a proven, scalable, high-performance data and
file management solution (based upon IBM General Parallel File System
or GPFS technology, also formerly known as code name Elastic Storage)
that's being used extensively across multiple industries worldwide.
2.1.4 GlusterFS
GlusterFS is a scalable network file system suitable for data
-intensive tasks such as cloud storage and media streaming. GlusterFS
is free and open source software and can utilize common off-the-shelf
hardware.
2.1.5 Swiftstack
SwiftStack provides an enterprise-grade object storage system and an
innovative storage controller that makes it simple for you to deploy,
integrate and manage object storage clusters in your data centers.
2.1.6 NexentaStor
NexentaStor is our flagship Open Source-driven SDS (OpenSDS) platform,
allowing thousands of customers all around the world to evolve their
storage infrastructure, increase flexibility and agility, simplify
management and dramatically reduce costs without compromising on
availability, reliability or functionality.
2.1.7 SUSE Enterprise Storage
SUSE Enterprise Storage is a highly scalable and resilient software
-based storage solution, powered by Ceph. It enables organizations to
build cost-efficient and highly scalable storage using commodity, off
-the-shelf servers and disk drives.
2.1.8 yStor
yStor is a software-defined solution that allows you to build an
enterprise-ready highly-scalable and distributed storage platform.
yStor provides elastic provisioning and unmatched flexibility without
the need for additional licenses, SAN hardware, or expensive
infrastructure components
2.1.9 SANsymphony-V
SANsymphony-V10 software is a comprehensive and scalable storage
services platform designed to maximize the performance, availability
and utilization of your IT assets, no matter how diverse they may be,
or what topology chosen.
2.1.10 VmWare (Virtual SAN)
VMware Virtual SAN is a radically simple, enterprise-class shared
storage solution for hyper-converged infrastructure optimized for
vSphere virtual machines.
2.1.11 Maxta (MxSP)
The Maxta Storage Platform (MxSP) provides organizations the choice
to hyper-converge on any x86 server, the ability to run on any
compute abstraction layer, and the flexibility to support any
combination of storage devices eliminating the need for complex and
expensive NAS and SAN devices.
2.1.12 Scality RING
The Scality RING is a proven software storage solution that enables
customers to build petabyte scale storage infrastructures leveraging
industry-standard servers.
2.1.13 Solution Comparison
Table 1 SDS Solution Features Comparison
--------------------------------------------------------------
| |Auto |Cent|Hetero- |Sto |Scale-|Elas |Self- |Tiered|
|Solu- |mated |ral |geneous |rage|in/out|tic/ |ser | |
|tions |Policy|ized|hardware|Type| |Resil |vice | |
| |-based| | | | |ience | | |
--------------------------------------------------------------
|HP | | | | | | | | |
|Store- | | V | V |F,O | V | V | | |
|Virtua l| | | | | | | | |
|VSA | | | | | | | | |
--------------------------------------------------------------
|ViPR | V | V | |O,B | V | V | V | |
--------------------------------------------------------------
|IBM | | | | | | | | |
|Spectrum| | | |F,O | | | | |
|Scale | V | V | V |B | V | V | | V |
|(GPFS) | | | | | | | | |
--------------------------------------------------------------
|Gluster | | V | V | B | V | V | | |
|FS | | | | | | | | |
--------------------------------------------------------------
|Swift | | V | V |F,O | V | V | | |
|stack | | | | | | | | |
--------------------------------------------------------------
|Nexenta | | V | V |F,B | V | V | | |
|Stor | | | | | | | | |
--------------------------------------------------------------
|SUSE | | | | | | | | |
|Enter | V | | V |O,B | V | V | V | |
|prise | | | | | | | | |
|Storage | | | | | | | | |
--------------------------------------------------------------
|yStor | | V | | B | V | V | | V |
--------------------------------------------------------------
|SAN | | | | | | | | |
|symphony| | V | V | B | V | V | | V |
|-V | | | | | | | | |
--------------------------------------------------------------
|VmWare | | | | | | | | |
|(Virtual| V | V | V | B | V | V | V | |
| SAN) | | | | | | | | |
--------------------------------------------------------------
|Maxta | | | | | | | | |
|(MxSP) | V | V | V |F,B | V | | | |
--------------------------------------------------------------
|Scality | | | | | | | | |
|RING | | V | V |F,O | V | V | | |
--------------------------------------------------------------
2.2 Standardization activities to SDS
The standardization activities in various organizations are at an
early stage. They are establishing SDS related working groups and
define SDS technology. In particular, open-source based storage
technologies, integrated management tool, interoperability
standardization discussion is active in many organizations.
This section provides various standardization activities related to
SDS, which is brief summary of each organization's description.
2.2.1 DMTF (Distributed Management Task Force, Inc.)
Cloud Management Initiative (CLOUD) is working to address management
interoperability for cloud systems. The following DMTF working groups
produce the SDS related standards and technologies promoted by the
Cloud Management Initiative:
* Cloud Management Working Group (CMWG) has developed DMTF
specification entitled "Cloud Infrastructure Management Interface
(CIMI)." The CIMI specification describes the model and protocol
for management interactions between a Cloud Infrastructure as a
Service (IaaS) provider and the consumers of an IaaS service. The
basic resources of IaaS (machines, storage, and networks) are
modeled to provide consumer management access to an implementation
of IaaS and facilitate portability between cloud implementations
that support the specification.
* Cloud Auditing Data Federation Working Group (CADF) defines the
CADF standard, a full event model anyone can use to fill in the
essential data needed to certify, self-manage and self-audit
application security in cloud environments.
* Open Virtualization Working Group (OVF) produces the Open
Virtualization Format (OVF) standard, which provides the industry
with a standard packaging format for software solutions based on
virtual systems.
Open Software Defined Data Center Incubator (OSDDC) is aim to develop
standard architectures and definitions to describe the Software
Defined Data Center (SDDC). The incubator is developing SDDC usecases,
reference architectures and requirements for industry standardization.
In addition, various related activities are ongoing such as Common
Information Model (CIM), Configuration Management Database Federation
(CMDBf), Systems Management Architecture for Server Hardware (SMASH),
etc.
2.2.2 SNIA (Advanced Storage and Information Technology)
The standardization activities in SNIA are mainly performed in the
Cloud Storage Technical Work Group. The Cloud Storage TWG acts as the
primary technical entity for the SNIA to identify, develop, and
coordinate systems standards for Cloud Storage. This group aims to
produce a comprehensive set of specifications and drives consistency
of interface standards and messages across the various Cloud Storage
related efforts. Representatively, Cloud Storage TWG promotes cloud
storage adoption with open standards such as "Cloud Data Management
Interface (CDMI)". CDMI is an ISO/IEC standard that enables cloud
solution vendors to meet the growing need of interoperability for
data stored in the cloud. There are currently more than 20 products
that meet the CDMI specification. SDS requires a standardized storage
management interface, such as "Storage Management Initiative
Specification (SMI-S)" developed by Storage Management Initiative
Specification (SMI-S) Core TWG, in order to automate the management
of the storage resources and discovery of their capabilities for use
in various pools. Besides, Object Drive TWG, Disk Resource Management
TWG, etc. are collaborating for Software Defined Data Center
realization.
2.2.3 OASIS (Organization for the Advancement of Structured Information
Standards)
OASIS Topology and Orchestration Specification for Cloud Applications
(TOSCA) Technical Committee is enhancing the portability and
management of cloud applications and services across their lifecycle.
TOSCA standards aim to enable Software Defined Environments (SDEs) by
optimizing the underlying cloud infrastructure.
Cloud Application Management for Platforms (CAMP) Technical Committee
defines interfaces for self-service provisioning, monitoring, and
control. Based on REST, CAMP is expected to foster an ecosystem of
common tools, plugins, libraries and frameworks, which will allow
vendors to offer greater value-add.
3. SDS
3.1 Definition of SDS
3.1.1 Definition
The SDS can be a software-based model that detaches the storage
management from the physical storage hardware and transforms it into
a service. The SDS as a logical storage pool can provide an automated
and centralized management service to administrator. The SDS might
also decrease storage management complexities. Moreover, the SDS is
able to improve efficiency, scalability and flexibility using
commodity hardware as well as to provide a cost effective technology
using existing storage resource. In this SDS environment, users can
request to allocate storage to applications by their requirements.
---------------------------------------------------------------------
| Service |
| ------------ ------------ ------------ ------------ ----------- |
|| VM || VM || App || App || App ||
| ------------ ------------ ------------ ------------ ----------- |
---------------------------------------------------------------------
---------------------------------------------------------------------
| Virtual Storage Pool |
---------------------------------------------------------------------
---------------------------------------------------------------------
| Heteroheneous hardware |
---------------------------------------------------------------------
3.1.2 Objectives
The objectives of the SDS are to improve efficiency and to reduce
wasting cost of storage. A facility cost was a problem of traditional
storage environment. Service providers had to construct big storage
service to provide service because they couldn't expect how much
storage space was exactly needed. Additionally, the traditional
storage had some problem such as vendor lock-in, difficulty of scale
out storage, and so on. On the other hand, the SDS can provide
storage to application unit and be managed by single control point.
It can support block, file, and object type storage using existing
hardware infrastructure. Moreover, it is possible to scale out
storage nodes with system disruption.
Table 2 Traditional Storage and SDS Comparison
---------------------------------------------------------------------
| Features | Traditional Storage | Sftware Defined Storage |
---------------------------------------------------------------------
| Scale up | Deploy entire | Add new shelves of disks |
| | new shelves of disks | |
---------------------------------------------------------------------
| Lock-in | Use specialized | Use heterogeneous hardware |
| | hardware | |
---------------------------------------------------------------------
| Service unit | System component unit | Application component unit |
| | | (Individual Container) |
---------------------------------------------------------------------
| Flexibility | Static | Dynamic |
| | | |
---------------------------------------------------------------------
| Management | Storage administrators' | Single control point |
| | intervention | |
---------------------------------------------------------------------
3.1.3 Benefits
* Simplicity
- Automated policy-based: To store data in the right place, at the
right time, with the right performance, and at the right cost
based on defined policies.
- Centralized management: To convert the whole storage hardware in
a seamless storage pool and offer a single control point to
manage all the resources
* Flexibility
- Heterogeneous hardware: To allow the use of commodity hardware
and the implementation over an existing infrastructure
- Supporting block, file and object storage: Integration of the
three basic storage types
* Scalability
- Scale-out architecture: To incorporate more storage nodes to
increase the capacity and improve the performance
- Elastic: Enable increase capacity as needed without disruption
the availability or performance
* Efficiency
- Component unit: To make traditional, big storage pool into
application unit storage to reduce wasting expense
3.2 Overview of SDS
3.2.1 Key Characteristics
3.2.1.1 Automated Policy-Based Management
* Data Provisioning: Provide users with access to data and resources.
* Data Protection and Availability
- Backup/Recovery: backup the data and restore it from backups.
- Snapshots: copy of the state of a system or to a capability
provided by certain systems at a particular point in time.
- Replication: the same data is stored on multiple storage devices,
to improve reliability, fault-tolerance, or accessibility.
- Clustering: linking many computers together to act like a single
computer and all de computers has access to all the data.
- Mirroring: the act of copying data from one location to another
in real time.
- Self-healing: the ability to perceive that it is not operating
correctly and make the necessary adjustments to restore itself to
normal operation.
- Data migration: Transfer data between storages.
- Data Performance
- Caching: a component that stores data so future requests for that
data can be served faster.
- Thin provisioning: a virtual provisioning mechanism that allows
addressable storage capacity to be provisioned without consuming
or reserving physical capacity.
- Auto-tiering: put data in the appropriate class of storage based
on how frequently the data is accessed in real-time.
* Event management and alerting
* Thin Provisioning
- This virtual disk form is very similar to the traditional format
with the exception that it does not pre-allocate the capacity of
the virtual disk from the datastore when it is created. When
storage capacity is required the virtual disk will allocate
storage in chunks equal to the size of the file system block.
3.2.1.2 Centralized Management
* Pool Management: create, deliver and manage storage pools.
* New Resources: manage new resources.
* Policy Settings: allow administrators to set policy for automate
managing the storage and data services.
* Service Levels: determine or set service level to the system
resources.
* Monitoring:
- Track capacity consumption.
- System Health.
- Monitor and report on performance trends from host to storage.
- View the physical resources and relationship dependencies.
- Event logs and system alerts management.
- Reports
* Troubleshoot: systematic search for the source of a problem so that
it can be solved, and so the product or process can be made
operational again.
3.2.1.3 Self-Service
* Users can easily subscribe to storage resources that meet their
workload demands.
* Users are not required to know or care about the underlying
hardware and software that is providing the storage to their
application.
* Automatically provisions the right hardware and software to meet
the users' needs based on policies pre-defined by the storage
administrator.
3.2.1.4 Heterogeneous Hardware/Storage Type
* Support Heterogamous Hardware as follows:
- Direct-attached storage (DAS)
- Network-attached storage (NAS)
- Storage area networks (SAN)
- Cloud Storage
* Support Block/File/Object Storage
- Block-based storage stores data on a hard disk as a sequence of
bits or bytes of a fixed size or length (a block).
- File-based storage systems, such as NAS appliances, are often
referred to as "filers" and store data on a hard disk as files
in a directory structure
- Object-based storage systems use containers to store data known
as objects in a flat address space instead of the hierarchical,
directory-based file systems that are common in block- and file
-based storage systems.
* Support implementation in virtual, physical environments or a mix
of them.
3.2.1.5 Scale-out Architecture
* Distributed computing
* Active-Active Architecture
* The data and the metadata are distributed across all nodes
* Allow Scale-Up
3.2.1.6 Scalability
* Scale while continuing to manage storage as a single enterprise
-class storage system.
* Provides massive, virtually limitless scalability.
* Scale without disruption the availability or performance.
4. References
[1] DMTF DSP-IS0501, Software Defined Data Center (SDDC) Definition,
2015
[2] SNIA, Software Defined Storage (White Paper), 2015
[3] VMware, The VMware Perspective on Software-Defined Storage (White
Paper), 2014
[4] EMC, Transform Your Storage For The Software Defined Data Center
(White Paper), 2015
Appendix A. Acknowledgements
This draft is supported by Institute for Information & communications
Technology Promotion(IITP) grant funded by the Korea government(MSIP)
(R-20150223-000247, Cloud Storage Brokering Technology for Data-Centric
Computing Standardization)
Authors' Addresses
Eui-Nam Huh
Computer Science and Engineering Department, Kyung Hee University
Yongin, South Korea
Phone: +82 (0)31 201 3778
Email: johnhuh@khu.ac.kr
Ga-Won Lee
Computer Science and Engineering Department, Kyung Hee University
Yongin, South Korea
Phone: +82 (0)31 201 2454
Email: gawon@khu.ac.kr
Yunkon Kim
Computer Science and Engineering Department, Kyung Hee University
Yongin, South Korea
Phone: +82 (0)31 201 2454
Email: ykkim@khu.ac.kr
Jintaek Kim
Consortium of Cloud Computing Research, Seoul, South Korea
Phone: +82 (0)2 2052 0156
Email: jtkim@cccr.ir.kr