In hopes of seeing some discussion kick off here, I'm very interested to hear more about the ECS solution. Mr. Hamilton, any opening remarks you have on where ECS can really show it's strength? Along those lines, it may be helpful for some short opening remarks from everyone to kick start the discussion?
Sure thing! First, let me say that it's exciting to see ECS Appliance, along with EMC Isilon, as a core part of the Data Lake Foundation Strategy announced at Strata + Hadoop World conference today.
ECS is the acronym for Elastic Cloud Storage. The ECS Appliance is a software-defined cloud-storage platform that combines the cost advantages of commodity infrastructure with the reliability, availability and serviceability (RAS) of an integrated solution.
ECS is an object-based storage platform so that's where it really shows its strengths. ECS is designed for REST-based applications and environments that have multiple sites and need massive scale. ECS features an active-active, read-write and and very efficient geo-replication that delivers low storage overhead without compromising access to data - especially for geo-distributed environments.
ECS supports industry leading Object and HDFS access protocols to support a broad range of cloud and Big Data applications. ECS also features seamless management, multi-tenancy, and detailed metering to deploy and manage storage-as-a-service. You can certainly find out more by visiting emc.com/ecsappliance.
Thanks, I'm looking forward to this Ask the Expert Event!
You know the latest OneFS 7.2.0 version quite well. In your opinion, what improvements or new features (since 7.1.1) are the most significant improvements and why? How does that impact customers?
OneFS 7.2 brings in so many new features. We introduced support for a new HD400 platform. As part of EMC's Data Lake Foundation, OneFS 7.2 includes support for HDFS 2.3, 2.4 and Apache Ambari integration for ease of provisioning, monitoring and managing Hadoop clusters.
We've enhanced our NFS stack to support NFS Access Zones and enable auditing capability on NFS.
OneFS 7.2 enables node compatibility for S210/S200 and X410/X400.
On the networking side, we've introduced support for source based routing.
A lot of feature enhancements in core OneFS to improve Snapshot performance, specifically snapshot delete performance.
Lastly, OneFS 7.2 now allows customers object access to OneFS using Openstack SWIFT.
If i want a customer, my first questions about the node would be
what is the performance of node?
there was a document with that information
the document was "Isilon IQ Clustered Storage for high-performance applications" created august 2008
do you have that document updated until onefs7.2 with the new models of nodes? and could you send me a copy?
the document show the values of performance with the protocols nfs, cifs, http and ftp, with the variables uncoalesced write, coalesced write, uncache read, cache read
that information would be vital to create new solutions for the client
it seems that the document you are referring to is no longer updated and released. There are several documents on HD400 and its specs available in the HD400 - Isilon Info Hub. Additional Info is also available on OneFS 7.2 here: OneFS 7.2 Documentation - Isilon Info Hub
I hope these libraries will provide you with the information you are looking for. If by any chance you are still missing specific documentation, please submit a request to firstname.lastname@example.org
HD400 has the lowest total cost of ownership per gigabyte than any product available to date. Because of the density of the HD400, as you scale out your cluster, your cost per gigabyte, and your cost per watt, continues to go down.
The HD400 is also more reliable than any platform Isilon has delivered to date. The HD400 leverages the improvements made to our filesystem journal NVRAM introduced with the X410 and S210. This NVRAM can ensure that any data is retained, even if a cluster experiences a power event that lasts months or years.
The HD400 is the most serviceable platform Isilon has deployed to date. The drives can all be serviced by pulling the drawer out. The fans and and power supplies are hot swappable. In the event of a cpu module failure, the 'suitcase' model of service allows a technician to service the cpu module and I/O components without taking the entire HD400 chassis out of the rack.
it seems that this document has not been updated since 2009. Instead many new document have been created, including the Isilon sizing tool for partners that can be used for performance indicators. I recommend you log in to the partner Portal and check out the Isilon sizing tool. If you are not familiar with it's location, this discussion thread has got directions to finding it.
Hope that helps,