Modern storage for modern business: get to know Hedvig
•Download as PPTX, PDF•
3 likes•859 views
Eric Carter, Senior Director of Marketing, on the rate at which data is growing and Hedvig's scalable solution to meet enterprise needs. Slides are from the Santa Clara University's Annual Conference on Massive Storage Systems and Technology 2015
Recommended
More Related Content
What's hot
What's hot (20)
Viewers also liked
Viewers also liked (18)
Similar to Modern storage for modern business: get to know Hedvig
Similar to Modern storage for modern business: get to know Hedvig (20)
Recently uploaded
Recently uploaded (20)
Modern storage for modern business: get to know Hedvig
- 1. 1 Copyright 2015 Hedvig Inc. Modern Storage for Modern Business Eric Carter Senior Director, Marketing Hedvig Inc. @ercarter
- 2. 2 Copyright 2015 Hedvig Inc. Business innovation requires flexibility Business innovation Time-to-market Flexible infrastructure Someone has a great idea Someone has to develop it Someone has to deploy it and manage it
- 3. 3 Copyright 2015 Hedvig Inc. According to Forrester . . . 3 10Xfaster growth of data than storage budgets 58%take days, weeks, or months to provision storage 14%have “cloud-like” storage provisioning Source: Forrester Technology Adoption Profile: Meet Evolving Business Demands With Software-Defined Storage, March 2015. Visit hedviginc.com for full research report.
- 4. 4 Copyright 2015 Hedvig Inc. The new way of doing storage Commodity Servers Software Software-Defined Storage + =
- 5. 5 Copyright 2015 Hedvig Inc. It’s not so new (as this audience knows…) …but it is new for most enterprises
- 6. 6 Copyright 2015 Hedvig Inc. 2012-present 2007-2011 2004-2007 It’s what Hedvig brings to the enterprise Commodity Servers Software Hedvig Distributed Storage Platform + =
- 7. 7 Copyright 2015 Hedvig Inc. ..and scales to massive proportion Elastic: Scale to petabytes of data Simple: Provide block, file, and object storage Flexible: Connect to any compute, or cloud
- 8. 8 Copyright 2015 Hedvig Inc. DC1 DC2 Cloud3 How it works iSCSI NFS 1 Create Virtual Disk(s) and assign policies via UI, CLI or API 2 Present block, file, or object storage to application tier 3 Capture I/O and communicate it to underlying cluster 4 Distribute and replicate data across cluster 5 Auto-tiers and auto-balance across racks, datacenters, and clouds Hedvig storage cluster Application hosts VM VM VM VM VM VM S3, Swift
- 9. 9 Copyright 2015 Hedvig Inc. Why do it this way? Copyright 2015 Hedvig Inc.9 Scale Seamlessly add or subtract capacity/performance – grow to 1000+ nodes Simplicity Simplify provisioning with storage resource abstraction – a few clicks Resilience Self-heal – never manually restore or migrate data again Leverage compute cluster, flash – ride commodity hardware power curvePerformance Cost Take advantage of low-cost whitebox hardware and appreciating software Flexibility Choose disk type and deployment – hyperscale or hyperconverged
- 10. 10 Copyright 2015 Hedvig Inc. Scale-out seamlessly with x86 or ARM Provision with unprecedented flexibility and speed Consolidate all protocols in one platform Run agnostic to any hypervisor, container, or OS Provide hybrid-aware DR, across any public cloud Support hyperconverged and hyperscale Enable enterprise-grade storage features Hedvig design principles
- 11. 11 Copyright 2015 Hedvig Inc. Who is doing it with Hedvig ;) Big data Quick, reliable indexing of 100M active client docs. Hedvig a scale-out, flash-friendly solution to replace local SSDs. Server virtualization Bringing apps back in house on virtualized, multi-datacenter architecture. Hedvig provides enterprise features and built-in resiliency. Private cloud Growing at a rate of 1 terabyte per hour in private and hybrid cloud environments. Hedvig provides scalability and multi- protocol flexibility for managed services.
- 12. 12 Copyright 2015 Hedvig Inc. Thank you!
Editor's Notes
- I’ll start with a quick acknowledgement of a dynamic that we see in the marketplace – and that is there is a desire to drive innovation by business executives – and this in today’s world rides atop technology moreso than at any other time in our history. This means there is pressure on application developers to develop apps in response to a business initiative and to get them to market quickly.
- According to Forrester, data is growing 10 times faster than are storage budgets. This creates a big challenge – how do you keep up? It also takes a very long time to provision new storage into an environment – 58% of those surveyed report that it takes days, weeks or even months to rollout a new solution. The bad news as well, is that while many are on the quest to gain a cloud-like capability, only 14% report having anything that even resembles the simplicity of cloud-like storage provisioning.
- More and more companies are moving away from traditional monolithic arrays to store VMs to a software-based approach – where the capabilities and value enterprise storage is purchased and delivered via software as opposed to locked up in a custom hardware solution. An WHY are companies considering and already adopting this kind of approach? One reason is Flexibility – today’s software-defined solutions – like Hedvig – give IT organizations the ability to zig and zag much more effectively with the demands of business. A recent Forrester study indicates that nearly 60% of businesses reports that storage provisioning takes days or longer with traditional solutions. With software defined, we’re in an era where provisioning is much more cloud-like, taking minutes to achieve versus hours, days or weeks. It also provides the ability to define virtual volumes with features and functions – like VMware VVOLS – assigning capabilities that uniquely fit a given application or service. A software approach also gives you the ability to bring your own hardware – to virtualize and pool your resources in ways that weren’t possible before – helping deliver significant value from existing and new assets.
- more and more companies are taking a page out of the playbook of the Googles, Amazons, and Facebooks, choosing to deploy off-the-shelf systems versus custom hardware. But why? Yes it is a cost thing – but it’s lower cost without sacrificing quality. We had a bit of a mild online debate recently as to the meaning of commodity. For many of us who have been in the storage industry for some time, it’s a foreign proposition – however Commodity doesn’t mean cheap and low quality, instead it signals that pound for pound there is parity across the market in terms of the components that are available. This is really where we are at with the server market. The quality of a given commodity node may differ slightly, but it is essentially uniform across producers. X86 servers, ARM servers, SSDs, and certainly hard drives. I argued in a recent block that if you pull the bezel on some of todays more traditional storage array solutions, inside you’ll see the same components that are available to you in the server hardware available from your provider of choice – Supermicro, Dell, Cisco, HP, etc. Paired with the software approach to storage, this means that for your virtualization environment, you can ride Moore’s law and take advantage of the frequent increases in processing power, drive capacity, etc. while maintaining or even dropping your costs over time. Commodity systems for many are easier to obtain and stock. One of our customers, chose software defined running on standard Cisco servers for their storage solution for this very reason. For them, it was much faster, and easier to obtain and deploy commodity Cisco servers to add capacity to their storage system than it was to try to requisition proprietary array hardware. IT also gives them the ability to get out of the game of pre-buying capacity in the traditional fashion in favor of a more just in time approach.
- At its most basic level, if you look at the equation on the screen – you take commodity servers - X86 or ARM – preferably packed with disk – including flash which we’ll talk further about shortly – and you add the Hedvig software – and you have what we call the Hedvig Distributed Storage Platform. Think about this as an elastic storage cluster to support all of your various application requirements. You bring to the party the kind of servers you’re comfortable with – typically off-the-shelf type hardware – and you leverage this well-known, easy-to-acquire hardware to run our software and support your storage needs. You determine the characteristics of your cluster – the quality of the servers you bring will contribute to the performance you can achieve with IOPs and so-forth – and the quantity of nodes contributes to the availability.
- And with Hedvig – this is what you can achieve - you simply add nodes to scale – even to petabytes of data. Hedvig lends itself very well to elastic workloads. When we set out to build the platform we knew going in that you have requirements for storage of various type – block, file and object- to enable the choice for you to deliver storage as you need it – all from the same platform. And with the flexibility to connect to any app, or compute platform, or even cloud that you want to take advantage of – along with a wide range of enterprise features we’ll discuss in the coming slides. Hedvig is a single solution you can leverage to meet your wide-ranging needs without having to resort to standing up islands of storage, train people on different products and figure out how to work with different vendors. Hedvig is all about simplicity and flexibility at scale.
- Now let’s take a look at a day-in –the-life if you will. Step one – an admin – or user – creates the Virtual Disk resource – this can be via UI, or CLI – or via API which some customers use to build their own self-service interface. Next that virtual disk is presented to a host via the Storage Proxy – providing a block of file resource – and in the case of object storage via S3 or Swift API. As hosts write data, the proxy intercepts the I/O and sends it down to the storage cluster. In the distributed systems approach, this data is distributed and stored in smaller chunks and tracked throughout the cluster – taking advantage of the horsepower and resources of the underlying x86 systems. Data replicas – the default being three – are then made and distributed as well – this is what allows for redundancy and high availability of data. And in the normal course of operations, the Storage Service will auto-tier data, auto balance across our server resources, and make sure that data is distributed to datacenters or clouds as dictated by your set policy. The great news here is that with Hedvig, you don’t need to be an expert in distributed systems to operate your storage system. The intelligence is built-into the software.
- This means utilizing storage software, and commodity hardware to deploy storage that takes advantage of many systems versus just a single box or a few – to scale out, not up – taking advantage of the power of many systems working together to deliver horsepower and capacity – an approach that takes advantage of parallelism. You can provision storage very quickly – very cloud like – with a few clicks. . And by many I mean, hundreds or more. A software approach also lends itself to flexible deployment models – running applications and storage services together in a hyperconverged fashion, or, decoupling and scaling compute and storage individually – what we refer to as hyperscale or a two-tier model. You also get the benefit of hardware advancements as they emerge – riding Moore’s law. Again from the Cloud-company playbook what we can achieve with a distributed storage system is a new level of resilience and performance. And imagine, removing old nodes and adding new nodes over time – and having data auto replicate and auto rebalance. You can finally get out of the business of doing data migrations and forklift system upgrades.With software – we also benefit from updates and enhancements over time such that it is an appreciating asset. This moves us away from the throwing hardware/software in the rubbish every 3-5 years and starting again. The software you invest in today can still be adding value to you on updated hardware for many many years. And programmability. Now this is NOT exclusive to software-defined– but software-based storage that is not locked into custom hardware, lends itself much more readily to cloud approaches, to microservices, and the like. For DevOps, software-defined storage is a must have, enabling automation for provisioning, scaling, VM and data movement. This approach helps you control costs – your investment in Hedvig software is an appreciating asset, protecting against technical and functional obsolescence.
- What are the use cases? First – we see server virtualization environments as prime places for you to leverage the Hedvig software-defined approach – especially where you are using multiple hypervisors. Where you need scalability, elasticity, and agility to keep up with all of these apps, span multiple locations and clouds, and accommodate the rate of change – Hedvig is a good fit. Next – for those building private clouds and looking to modernize how you deliver IT, you need a storage platform that is designed to scale with cloud-like capabilities. Hedvig provides the flexibility to support a wide range of options for storage configuration and for scalability, taking advantage of commodity economics. With Hedvig, you can achieve a managed service offering that provides onsite and offsite storage capabilities with built-in disaster-recovery-as-a-service. For big data Hadoop deployments – Hedvig provides a single, efficient storage pool – what the industry is now calling a data lake – to deliver data into multiple Hadoop distributions. Hedvig delivers tunable capabilities and features like dedupe and client-side caching to deliver efficiency and performance for big data.