Presentation is loading. Please wait.

Presentation is loading. Please wait.

Technical Training Hadoop 101

Similar presentations


Presentation on theme: "Technical Training Hadoop 101"— Presentation transcript:

1 Technical Training Hadoop 101
November 2016

2 Isilon Momentum 7,000+ Customers World Wide
1,300 New Customers in 2015 #1 in Data Lakes >1,200 HDFS Customers >12,000 Clusters in the Field Approaching 1PB Average Cluster Capacity Scale-Out NAS Leader

3 Hadoop Market Leadership
Market Leader in Hadoop Shared Storage #1 Rapidly growing use case Customers 1200+ Pivotal Cloudera IBM Isilon was the first scale out NAS product to natively integrate HDFS. We did that over two years back. Over the last couple of years, we have grown to become the #1 market leader in Hadoop Shared Storage. We have over customers and are growing at over 250%. We work with all leading commercial apache Hadoop distributions including Pivotal and Cloudera. We are still in the early stages of this industry. We believe that this workload will continue to grow as business begin to realize the rich value of data and the insights they can glean from it to drive their bottom-line. So what are we doing to help our customers unlock the value of their data… (click)

4 Isilon = Enterprise Open Source Hadoop
Benefits Deployment Enterprise Proven Security and Data Protection Separate Compute and Storage Cost Effective Scaling Lower TCO—Management, Power, Cooling, Footprint Multi-protocol—Bring the Analytics to the Data Reduced Time to Results with Consistent SLA’s Isilon supports the HDFS interfaces for the DataNode and NameNode to host data and metadata Underlying file system is OneFS As simple as pointing the HDFS clients to the DNS name of the Isilon cluster!

5 Support for Standard Apache Components
Isilon’s HDFS implementation supports standard Apache Hadoop RPC protocols and commands Isilon is 100% compatible with Apache compliant Hadoop distributions Components & Applications that run on Apache Hadoop run seamlessly on Isilon

6 ONEFS OPERATING SYSTEM
Single File System One Namespace Simplicity & Ease of Use High Performance Linear Scalability The Isilon OneFS operating system provides the intelligence behind all Isilon scale-out storage systems. It combines the three layers of traditional storage architectures—file system, volume manager, and data protection—into one unified software layer, creating a single intelligent file system that spans all nodes within an Isilon cluster. OneFS provides a number of important advantages: A single file system for simplicity and great ease of management Leading NAS storage performance (based on SPECsfs 2008 benchmark for CIFS environment) Unmatched efficiency with over 80 percent storage utilization plus automated storage tiering to gain additional efficiencies Easy, “grow as you go” flexibility Linear scalability lets you can scale performance and capacity up to 50 PB in a single Isilon cluster Unmatched Efficiency Easy Growth 6

7 ENTERPRISE GRADE SOFTWARE
DATA PROTECTION & EFFICIENCY DATA MANAGEMENT ENTERPRISE GRADE SOFTWARE DATA MANAGEMENT SnapshotIQ Fast, Efficient Data Backup And Recovery SyncIQ Fast And Flexible Asynchronous Replication For Disaster Recovery Protection SmartConnect Policy-based Client Failover With Load Balancing SmartLock Policy-based Compliance and WORM Data Protection SmartDedupe Data Deduplication to reduce storage requirements and costs SmartPools Policy-based Automated Tiering SmartQuotas Quota Management And Thin Provisioning InsightIQ Performance Monitoring And Reporting To Manage Storage Resources CloudPools Cloud-scale Capacity Data lake has enterprise grade features that strengthen it. It include features on Data Protection and Data management Discuss any of the features here…

8 HDFS Completely Rewritten in OneFS 8.0
HDFS protocol rewritten in C++ Increased parallel processing Greater scalability Support for audit, CloudPools, and SMB file filtering New web administration interface Full configuration options in web administration interface Improved CLI options isi hdfs command controls HDFS settings In OneFS 8.0, the Isilon engineering team made the decision to provide a robust and scalable version of HDFS for this and all future releases. Starting in OneFS 8.0, the HDFS protocol was entirely rewritten in C++ code to increase processing, scalability, a web administration interface, as well as to add additional support for auditing, CloudPools, and SMB file filtering. With this rewrite OneFS 8.0 has a new foundation, purpose built, to support continued future HDFS innovations.

9 ONEFS 8.0 HDFS LOAD BALANCING
(1) SmartConnect DNS Policy Load Balancing HDFS Namenode Traffic Hadoop Compute Node DFSClient HDFS Datanode Traffic (3) Read HDFS Blocks (possible Adaptive Prefetch) (4) Write HDFS Blocks (2) Namenode Session: addBlock (Write) getBlockLocations (Read) Virtual Rack Isilon Node Isilon Node Isilon Node Isilon Node Smart Connect Svc.IP Namenode Namenode Namenode Namenode Datanode Datanode Datanode Datanode Single Isilon Cluster (Infiniband Backend)

10 Onefs 8.0 Datanode load balancing
Intelligently improve your hadoop performance Key Features HDFS Client Intelligently provides datanode with the least load to new HDFS clients Totally transparent to client, no configuration required 1. Namenode: Where to write? 3. Good, will write to Node 2. Benefits 2. Write to Node 2. Improves overall performance of Hadoop clients for analytics workloads Avoids overloading any specific OneFS node and increases cluster resilience Node 1 Detailed Speaker Notes Node 2 Node 3 Connection Count

11 Onefs 8 HDFS Web Interface
New HDFS configuration page in web administration interface (OneFS 8) Can enable HDFS and change block size Authentication type and root directory: Any configuration previously done via CLI now done in web administration interface Previously, Hadoop could only be configured through the CLI. Now with OneFS 8.0, administrators have an HDFS page in the web administration interface that allows them to configure all the settings and clients through either the CLI or the web administration interface depending on their preference. All configuration options are available through either interface.

12 Onefs Access Zones An access zone is:
A way to carve the cluster into smaller clusters A way to control access based on individual authentication OneFS’s Multi-Tenancy solution Access Zone-1 Domain Controller-1 Kerberos-1 Group Database - 2 System Zone Group Database - 1 Kerberos-2 Access Zones allow us to carve the cluster up into smaller clusters. In prior versions of OneFS, only SMB was zone aware, meaning that the SMB clients could authenticate through individual access zones while all NFS clients had to authenticate through the System zone as NFS was not zone aware. Now with OneFS v7.2 Access zones are now NFS aware, meaning the exports and aliases can exist and be visible on a per zone basis Isilon provides secure multi-tenancy with access zones with all protocols. Access zones do not require a separate license. Access zones enable you to partition cluster access and allocate resources to self-contained units, providing a shared tenant environment. You can configure each access zone with its own set of authentication providers, user mapping rules, and SMB shares or NFS exports. An access zone is a context that you can set up to control access based on a connecting client IP address. The system zone is typically reserved for administration use. The purpose of an access zone is to define a list of authentication providers, such as Domain Controllers, group databases, LDAP, NIS or Kerberos Servers, that apply only in the context of the zone you created. Lets use the analogy of a maitre d’ at a fine restaurant The domain controller is the maitre d’ checking your reservations at Chez SMB restaurant similar to the Kerberos Servers typically performing the same function but at Chez NFS. Other authentication mechanisms, like group databases, LDAP and NIS play a similar role. Whether you are using Domain controllers or Kerberos servers they manage access and authentication. All user access to the cluster is controlled through access zones. Each access zone contains all of the necessary configuration to support authentication and identity management services on OneFS. What does this all mean? Now your customer can provide multi-tenancy, or segregated access for separate departments within their corporation like Legal and Finance. This will provide complete segregation as if they each had their own cluster. And the best part is we now enable customers to support chargebacks so that you simply can parcel out the cost of the storage platform. Access Zone-2 Domain Controller-2 LDAP-1 NIS - 1

13 DNS per Access Zone Domain Name Resolution per Access Zone
Best fit in environment where each tenancy (group) has dedicated domain- name directory service Foundation piece to isolate client network connectivity associated with directory services .legal.bb.com Access Zone-1 Domain Controller-1 DNS1 Kerberos-1 LDAP-1 System Zone DNS0 Local Database-1 Before Riptide, Access Zones can be defined to provide separate data container with different authentication provider for each tenancy’s clients. In Riptide, Access Zone is further enhanced to allow each zone to be associated with a distinct DNS server, facilitating independent IP address resolution based on each tenant’s networking configuration. This is critical to environment where 1) DNS forwarding 2) a single external DNS for all tenancies are not acceptable to the storage owner. For use cases, this address the need to support storage consolidations or internal storage cloud. In either cases, it is very typical that all the tenancy zones are used to serve different organizations where both domain name and network configuration are totally independent and separate. Hence, DNS per Access Zone is a much desired feature in these environments. hr.bb.com Access Zone-2 Domain Controller-2 DNS2 NIS-1

14 Kerberos Security Enhancements
OneFS 8.0 enhanced Kerberos encryption support Add AES256 library Enables AES 128-bit and AES 256-bit encryption support Previous releases supported RC4 and DES encryption only Enabled by default No setup required to enable support Meets customer security requirements and expectations In pre-OneFS 8.0, OneFS supports RC4 and DES for Kerberos encryption only. RC4 and DES are types of encryption that are less secure than AES encryption. When a Kerberos encryption request was received, OneFS would auto-negotiate the encryption to either RC4 or DES depending on the ticketing service. In OneFS 8.0, support for AES 128-bit and AES 256-bit encryption was added for Kerberos using an AES256 library. AES encryption is enabled by support and is automatically selected as part of the Kerberos process when requested. No additional setup is required. AES encryption support enables Kerberos in OneFS 8.0 to meet customer required security levels.

15 Isilon Encrypted Clusters
Simplified Encryption using SED Transparent, Always On, Everything Encrypted No mixing of SED and non-SED drives or nodes Value No costly external equipment in addition to the storage you’re already deploying Maximize Performance Encryption workload is distributed to each drive Less than 1% impact compared to non-SED Internal Key Management Highly available, internal key management Key Message We have a simplified data-at-rest encryption solution using SEDs Starting July 2014, we now support self-encrypting HDDs and SSDs on all our platforms S210 (900GB HDD, 800GB SSD) (S200 not supported) X series (3TB or 4TB HDD, 800GB SSD) NL series (3TB or 4TB HDD; 800GB SSD only available by exception)

16 SED Availability Matrix
900GB SAS 3TB SATA 4TB SATA 6TB SATA* 8TB SATA* 800GB SSD 1.6TB SSD S200 S210 Y 2H2016 X400 X410 X200 X210 NL400 NL410 Y* 2H2016* HD400 Y (with 6TB) 2H2016 (with 8TB) * OneFS 8.0 and later only

17 HDFS Ambari integration
Key Features Hadoop Management made simple Leverage Ambari to monitor key performance metrics and alerts of OneFS file system Key metrics like disk, CPU, network and namenode usage are all reported Benefits Single management point for Amabri operator to manage a Hadoop cluster with OneFS Ambari admin can proactively monitor and troubleshoot performance issues with OneFS file system similar to DAS Detailed Speaker Notes

18 Authorize only the right users to access HDFS FILES
Ranger Integration Authorize only the right users to access HDFS FILES Key Features Ranger Enables Ranger authorization policies to be executed in OneFS OneFS native file access control continue to be effective Dual access control checks guarantee file access meet both Hadoop and IT admin needs Ranger Authorization Policies Benefits Ranger admin can enforce Hadoop access policies across all Hadoop components consistently IT/datacenter admin maintains their control on multi-protocol datalake access in OneFS OneFS Native Access Control Check Detailed Speaker Notes

19 Available now Available Now Isilon Hadoop Tools
Automate user/group creation for Hortonworks, Cloudera, PivotalHD, and IBM Big Insights Creates directory structures according to Hadoop distribution github.com/isilon/isilon_hadoop_tools

20 Isilon for Multiple Analytics Applications
NFS NFS SMB name node data node SMB SMB, NFS, HTTP, FTP, HDFS HDFS MAP Reduce With Isilon, the compute nodes can run a variety of Hadoop distributions including Apache, Pivotal, Cloudera, and Hortonworks. In fact, these can all run jobs simultaneously on the SAME data sets. This allows companies to easily try and switch to other distributions easily, without a complicated and costly data migration. NFS

21 ISILON SCALE-OUT ARCHITECTURE
Windows Web Ethernet Protocols REST SWIFT NFS HDFS NDMP HTTP SMB FTP SECURE Apps INTEGRITY CONFIDENTIALITY Mac/iOS AVAILABILITY Cloud Linux/Unix Let us take a look at the EMC Isilon scale-out NAS architecture: Starting at the Client/Application layer, the Isilon NAS architecture supports a wide range of operating system environments natively Isilon Multi-Protocol capabilities support NFS, CIFS, HTTP, FTP, HDFS for Hadoop and Data Analytics, and REST for Object and Cloud computing requirements. This provides you with great interoperability for business applications as well as your data analytics activities. At the Ethernet level, the Isilon supports redundant Gig-e 10 Gig-e Network providing the performance your workloads need OneFS is a single file system/single volume architecture, which makes it extremely easy to manage, regardless of the number of nodes in the storage cluster. Isilon storage systems scale from a minimum of 3 nodes up to 144 nodes, all of which are connected with an InfiniBand communications layer. This enables you to scale your data lake foundations from 18TB to over 50PB in a single volume. You can leverage record performance and density without compromising on simplicity. Archive Hadoop 10 Gig-e and IB (Today) 10 and 40 Gig-e (Gen 6 HW 1H17)

22 Isilon’s Strategy For Data Insights
Actionable Understanding Helping users understand their Isilon Environment Presenting the right information in a meaningfully, accessibly and understandably Improved storage utilization Capacity planning and forecast Track who’s using the most resources Personalized dynamic information Enabling development of tailored Data Insights

23 The Data Insights Application
Powerful cluster monitoring & reporting on Flexible reporting on file system characteristics View “top N” and drill into directory & file sizes, ages, access time, etc. Track health and performance of the cluster Capacity forecasting Estimate capacity utilization based on selected time frames Plan for cluster expansion or file system cleanup activities Live performance statistics for cluster, node, protocol, client, and more View and drill into throughput, operations, CPU utilization, etc. to plan optimization efforts Feature Benefit Powerful, “real-time” monitoring and reporting of Isilon clusters Track health and activity of clusters and determine usage and growth trends Capacity usage forecast tool Capacity planning for cluster (as well as tiers and pools within the cluster) “Real-time” performance stats including throughput, IOPS, and many more Understand “what’s going on” to inform optimization efforts Filter and breakout charts by client, protocol, and other drill-downs See “who’s doing what” to address user problems, and problem users Explore improved file system reports from OneFS 8.0 Deeper understanding of what content is where

24 Insightiq 4.0 Key Features
Supports OneFS 8.0 and IsilonSD Edge clusters New, easy to use capacity utilization forecast tool Improved performance and reliability to scale with data growth Added support for IPv6 reporting Explore fresher and more reliable File System Analyze data from OneFS

25 capacity forecast tool
Select the Range on which to base the forecast Dotted line shows predicted utilization Filter by tier or nodepool for even more meaningful forecasting Forecast date highlighted along with growth rate

26 CLOUDPOOLS CORE CLOUD PROVIDER
HOT DATA >30 days WARM DATA 1-2 Months FROZEN DATA 1-2 years So lets take a look at how Isilon CloudPools software work. In this example wo have an Isilon cluster with a set of high performance nodes: <click> and we can add an another tier of storage and transparently move older data to that less expensive tier recently we introduced the high capacity HD400 node for deep archive and we can use our policy engine to move data within our data center based on business rules For really old files, the storage owner may want to free up space on the primary storage and tier the data offsite to a cloud provider. To the end user, it still needs to appears that the data is resident on the data center. When users access data resident on the high performance tier, data is retrieved quickly. When users access data resident on the lesser performance tier in the data center, data is retrieved somewhat slowly. For infrequently accessed data that is on the cloud provider, data access is the slowest but it is transparent to users and apps.. 26

27 Seamless CLOUD INTEGRATION
CLOUDPOOLS CORE CLOUD PROVIDER Seamless CLOUD INTEGRATION APPS & USERS When users access data resident on the high performance tier, data is retrieved quickly. <click> When users access data resident on the lesser performance tier in the data center, data is retrieved somewhat slowly. For infrequently accessed data that is on the cloud provider, data access is the slowest but it is transparent to users and apps. Access time 27

28 CLOUDPOOLS ENABLES FLEXIBLE CHOICE
Private Hosted Public Great for Larger Datasets Smaller Datasets, Bursting Cost Lowest overall TCO Predictable Opex Variable Opex Latency Low Med-High High Data Residency Concerns None Compliance Control Low-Med PRIVATE DATA LAKE HOSTED Comparing the CloudPools options: Private Cloud Benefits (with ECS) Simple multi-purpose storage that provides infinite scalability with cloud- like economics Eliminates any data residency or sovereignty concerns for ‘no public cloud’ policy organizations Optimized performance with low latency with co-location of ECS and Isilon instances Hosted Cloud Benefits (with Virtustream) Enterprise cloud software that specializes in moving complex enterprise IT to the cloud Flexible cloud options – Public, Private or Hybrid solutions Run enterprise IT and mission critical applications in the cloud Public Cloud Benefits (with AWS/Microsoft Azure/etc.) Low operational cost Easy to deploy and easy to scale / add capacity (Public Cloud will take care of all the work) Perceived low initial storage cost PUBLIC

29 CLOUDPOOLS – PROXY SUPPORT
OneFS v 8.0.1 CLOUDPOOLS – PROXY SUPPORT Key Features EMC2 Isilon ECS Multiple Isilon nodes can now simultaneously tier to the cloud Ability to update proxy servers Internet Benefits Proxy No direct external network exposure of Isilon systems for CloudPools No network workarounds necessary to configure CloudPools EMC announces new infrastructure flexibility to use OneFS CloudPools tiering with proxy servers Today, enterprise customers deploy complex network topologies to protect their critical data and core infrastructures are deployed behind a complex web of firewalls and load balancers. Prior Isilon CloudPools feature will require either exposing a number of Isilon nodes to external network to facilitate multi-node tiering or require complex source routing rules to circumvent corporate proxy servers. Now, we have a simpler path forward Highlight CloudPools will support proxy servers. Multiple Isilon nodes can now simultaneously tier to the cloud without direct external network exposure or network workarounds.

30 Linear Scaling of Performance and Capacity
Dell emc ISILON FAMILY Linear Scaling of Performance and Capacity S-Series High Performance Platform X-Series Highly Versatile Platform Reduced Costs Performance Nearline Platform NL-Series High Density Platform HD-Series IsilonSD Software Defined Your hardware CloudPools Internal Cloud External Cloud The Isilon scale-out storage product family includes four flexible product lines tailored for specific business needs. You can choose a combination of these different platform nodes to create a storage solution that meets your specific needs: Note to Presenter: Click in Slide Show mode for animation.. S-Series: our platform for high transactional workloads The Isilon S-Series (S210) combines unmatched IOPS performance with high efficiency and an ultra-low overhead scale-out NAS package. (5.4TB to 28.8TB per node) With SSD technology for file-system metadata and file-based storage workflows, the S- Series delivers additional performance gains for metadata-intensive operations while improving overall latency. X-Series: The Isilon X-Series, our most flexible and comprehensive storage product line, strikes the right balance between large-capacity and high throughput performance storage. (X210: 7.2TB to 48TB per node) (X410: 36TB to 144TB per node) The highly versatile X-Series is an ideal solution for high-concurrent and sequential throughput applications. NL-Series: our platform for economical near-line storage The NL-Series (NL410) is designed to provide cost-effective, large-capacity storage. (36TB to 210 TB per node). The result is a highly economical, massively scalable storage solution at an extremely attractive price per terabyte of capacity and low overall TCO. The NL-Series is a great solution for large-scale data archiving. HD-Series: our new high density platform The Isilon HD-Series (HD400) provides highly efficient and resilient scale-out storage platform for unstructured data that can scale to over 50 PB in a single file system. With massive scalability, robust data protection options and a high density footprint that lowers operating costs by 50%, the HD-Series is ideal for Big Data storage needs including deep archiving solutions CloudPools is software that allows you to integrate Isilon with the cloud. Now, you can tier inactive data to your choice of Cloud providers IsilonSD Edge is SDS for the enterprise edge locations that runs on commodity hardware on top of Vmware Capacity

31 DELL EMC Isilon nitro All Flash, Scale Out NAS
Isilon’s highly dense, extremely modular and incredibly scalable all-flash tier Targeted at extreme performance NAS markets Start a cluster with a single chassis & scale to 100+ chassis is a single global namespace Get up to 1PB of Flash from 1 4U chassis All OneFS Features Supported!

32 100+ 400+ Petabytes NODES! Integrates into your existing cluster
SECURE INTEGRITY CONFIDENTIALITY AVAILABILITY REST SWIFT NFS HDFS NDMP HTTP SMB FTP #1 DIFFERENTIATOR Enterprise Features: SmartPools CloudPools SmartConnect SmartLock SmartQuotas SyncIQ SnapshotIQ InsightIQ Integrates into your existing cluster Single namespace, from flash to spinning drives to cloud

33 Nitro Use cases HIGH THROUGHPUT: large datasets of large files for parallel processing IOps INTENSIVE: Billions of Small file, large datasets for parallel processing PREDICTABLE LATENCY: Predictable performance for mixed workloads IMPROVED TCO: Relief from infrastructure and energy efficiency Lossless high quality media output Quickly finding outliers and variations – DNA sequencing Pattern and trends search - Weather data Content repository Compute intensive Big data analytics Enterprise applications Multimedia & content delivery Operational cost confined Energy efficiency mandates Infrastructure constrained Media & Entertainment Electronic Design Automation Life Sciences Geoseismic IoT Government High Performance Computing

34 Isilon platforms

35 S210 specifications Next generation of S-Series
CPU: dual, 6-Core Ivy Bridge Processors RAM: 32GB to 256GB Drives: 24 X 2.5-inch bays HDD: 2.5-inch SAS, 300GB-1.2TB each SSD: up to 6 SSDs, 200GB-800GB each Self-encrypted options available Front-end I/O: 2x1GbE + 2x10GbE Back-end I/O: QDR IB, 1m-100m cabling Chassis Standard 2U enclosure Dual redundant, hot swappable PSUs Key Message: S210 is a refresh of the S200 to provide even better performance Newer generation Intel CPUs for additional horsepower and future performance headroom Higher RAM (256 vs 92 in S200) for larger caching QDR IB for up to 100m cabling Additional details: CPU: dual, 6-core E5-2620v2 RAM: DDR3, 16 DIMM slots Cable Length: Planning to qualify up to 100m PSU: 875W power supplies with both 120V and 240V support Node equivalence with S200: Not available at Jaws launch. Customers need to continue purchasing S200. We will have node equivalence by end of year (with Moby) with a relaxed RAM constraint: The S210 can be the memory config that is 1 below or 1 above the S200 RAM. This is because the S200 and S210 offer memory in different increments; S200 uses 6 or 12 DIMMs, whereas S210 uses 8 or 16 DIMMs. E.g. you have a node pool with S GB RAM. You can add S210 with either 32GB or 64GB RAM. Drive (HDD and SSD) count and capacity must still match within a node pool.

36 X210 specifications Next generation of X-Series
CPU: Intel E5 2407v2 – 4 Cores RAM: 6 DDR3 slots 24GB - 48GB RAM Drives: 12 X 3.5-inch bays HDD: 3.5-inch SATA, 1-4TB each SSD: up to 6 SSDs, GB each Self-encrypted options available Front-end I/O: 2x1GbE + 2x10GbE Back-end I/O: QDR IB, 1m-100m cabling Chassis Standard 2U enclosure Dual redundant, hot swappable PSUs) 3.48” x 18.87” x 28.5” 61.1 lbs Key Message: S210 is a refresh of the S200 to provide even better performance Newer generation Intel CPUs for additional horsepower and future performance headroom Higher RAM (256 vs 92 in S200) for larger caching QDR IB for up to 100m cabling Additional details: CPU: dual, 6-core E5-2620v2 RAM: DDR3, 16 DIMM slots Cable Length: Planning to qualify up to 100m PSU: 875W power supplies with both 120V and 240V support Node equivalence with S200: Not available at Jaws launch. Customers need to continue purchasing S200. We will have node equivalence by end of year (with Moby) with a relaxed RAM constraint: The S210 can be the memory config that is 1 below or 1 above the S200 RAM. This is because the S200 and S210 offer memory in different increments; S200 uses 6 or 12 DIMMs, whereas S210 uses 8 or 16 DIMMs. E.g. you have a node pool with S GB RAM. You can add S210 with either 32GB or 64GB RAM. Drive (HDD and SSD) count and capacity must still match within a node pool.

37 X410 specifications High Throughput Platform
CPU: dual, 8-Core Ivy Bridge Processors RAM: 32GB to 256GB Drives: 36 X 3.5-inch bays HDD: 3.5-inch SATA, 1-4TB each SSD: up to 6 SSDs, GB each Self-encrypted options available Front-end I/O: 2x1GbE + 2x10GbE Back-end I/O: QDR IB, 1m-100m cabling Chassis Standard 4U enclosure Dual redundant, hot swappable PSUs (high line only) Key Message: X410 is a refresh of the X400 to provide higher throughput performance Newer generation CPU for performance now and in the future Larger RAM options (256 vs 192 in X400) for larger caches QDR IB for up to 100m cabling Note: X410 currently only available with 240V-compatible power supplies. Will be introducing 120V-compatible version later this year. Additional Details: CPU: dual, 8-core E5-2640v2 RAM: DDR3, 16 DIMM slots. Actual configurations are 32GB, 64GB, 128GB, or 256GB. Cable Length: Planning to qualify up to 100m PSU: 1100W power supplies. 240V only. We’re working to upgrade the PSU to one that also works with low line ( V) by end of This means that APAC customers will need to wait until end of 2014 to purchase low-line capable X410s. Node equivalence with X400: Not available at Jaws launch. Customers need to continue purchasing X400. We will have node equivalence by end of year (with Moby) with a relaxed RAM constraint: The X410 can be the memory config that is 1 below or 1 above the X400 RAM. This is because the X400 and X410 offer memory in different increments; X400 uses 6 or 12 DIMMs, whereas X410 uses 8 or 16 DIMMs. E.g. you have a node pool with X GB RAM. You can add X410 with either 32GB or 64GB RAM. Drive (HDD and SSD) count and capacity must still match within a node pool.

38 HD400 High Density Platform
FEATURES HD400 High Density Platform 2.5X Increase in cluster capacity to 67PB 464 TB per node and 4.2 PB per rack Pre-Racked Configuration Options KEY BENEFITS Massive scalability for unstructured data consolidation and storage 50% lower operational expenses Ideal for deep archiving, DR & Data Lake Foundation The new Isilon HD400 high density platform raises the bar significantly on data storage capacity in a simple-to-manage, scale-out platform: Key features include: 2.5X capacity increase from our maximum cluster capacity of 20 PB in to 50 PB today with the new Isilon HD400 high density platform Each HD400 node includes fifty-eight 8 TB drives in a 4U chassis for a total capacity of 464 TB/node This translates into a highly dense 4.2 PB Capacity / Rack (note: 9 x HD400 nodes per rack) Pre-Racked Configuration Options Key Benefits include: Massive scalability for unstructured data consolidation and storage 50% Lower Operational Expenses including Power, Cooling and Data Center Floor Space Ideal For Deep Archiving, Disaster Recovery target and Data Lake Foundation

39 HD400 Typical Use CaseS HD400 High Density Platform
Deep Archives: Large-scale, high density, deep archiving storage with unmatched efficiency to lower costs and robust data protection and security options Disaster Recovery: Highly efficient, large capacity disaster recovery target Data Lake Foundation: Combine with Isilon S- Series, X-Series, and NL-Series nodes to provide an efficient Big Data storage solution that supports a broad range of traditional and next-generation workloads with a single Isilon cluster The Isilon HD400 is designed to provide enterprises with a highly efficient and resilient scale-out storage platform for unstructured data that can scale to over 50 PB in a single file system. With massive scalability, robust data protection options and a high density footprint that lowers operating costs by 50%, the Isilon HD-Series is ideal for Big Data storage needs including deep archiving solutions. Typical Use case examples Deep archives: for large-scale, high density, deep archiving data storage that offers unmatched efficiency to lower costs and provides robust data protection and security options Disaster recovery: provides a highly efficient disaster recovery target for organizations requiring an economical, large-capacity storage solution Data lake foundation: can be combined with Isilon S-Series, X-Series, and NL-Series nodes to provide a highly efficient Big Data storage solution that can support a broad range of traditional and next-generation workloads with a single Isilon cluster

40 specifications NL410 Nearline Storage Platform
CPU: Intel E5 2407v2 – 4 Cores RAM: 6 DDR3 slots 24GB - 48GB RAM Drives: 36 x 3.5-inch bays HDD: 3.5-inch SATA, 1-8TB each SSD: 0,1 or 2 SSD, GB each Self-encrypted options available Front-end I/O: 2x1GbE + 2x10GbE Back-end I/O: QDR IB, 1m-100m cabling Chassis Standard 4U enclosure Dual redundant, hot swappable PSUs 6.96” x 18.90” x 31.25” 118 lbs / 54 kg The Isilon HD400 is designed to provide enterprises with a highly efficient and resilient scale-out storage platform for unstructured data that can scale to over 50 PB in a single file system. With massive scalability, robust data protection options and a high density footprint that lowers operating costs by 50%, the Isilon HD-Series is ideal for Big Data storage needs including deep archiving solutions. Typical Use case examples Deep archives: for large-scale, high density, deep archiving data storage that offers unmatched efficiency to lower costs and provides robust data protection and security options Disaster recovery: provides a highly efficient disaster recovery target for organizations requiring an economical, large-capacity storage solution Data lake foundation: can be combined with Isilon S-Series, X-Series, and NL-Series nodes to provide a highly efficient Big Data storage solution that can support a broad range of traditional and next-generation workloads with a single Isilon cluster

41


Download ppt "Technical Training Hadoop 101"

Similar presentations


Ads by Google