Welcome to this Dell EMC Community Network Ask the Expert conversation. Our discussion is now open for interaction.
When formulating your question please consider the following:
Login or register to engage – You must login or, if you’re new, register to DECN in order to post a question, like, bookmark or follow this discussion. https://developer-content.emc.com/login/login.asp.
Be respectful – SMEs are volunteering to assist you and will provide you with information that they have access to.
Provide details and background – Our SMEs will try their best to provide a full answer for your satisfaction, but will appreciate as much information you can provide about your inquiry.
Keep focus and within topic – The expectation is to receive questions about the topic is that is being presented on the title and description of this event. We would appreciate if questions are kept within those boundaries.
Be appreciative – Many tools available for this, likes, user ratings, follow, bookmark and the good old “thank you”.
You have until November 2nd to post questions.
We’re delighted to know you’ve decided to join our debate and hope you receive the answers you’re eager to find. Thanks!
The new Isilon All-Flash platform will provide blazing performance to run a wide range of the most demanding NAS workloads. Examples include:
1) Data Lake: Customers looking to consolidate their unstructured data with a data lake that supports a wide range of workloads including a number of applications that require extreme NAS performance.
2) Media & Entertainment: M&E customer target use cases include: 4K streaming of data to drive higher resolution of viewing content; and large-scale render farms that need to process larger datasets and generate content faster.
3) Life Sciences: Target use cases include genomics sequencing and large-scale microscopy environments that need to process enormous datasets in a short time period
4) Financial Services: Use cases will include workloads that need high performance, low latency and maximum throughput such as ticker analysis that requires near real-time analysis of stock market data on very large datasets.
5) Big Data Analytics: Customers looking to run a wide range of near real-time analytics with their choice of Big Data analytics vendors like Hortonworks, Cloudera, Pivotal, IBM and Splunk.
6) Big Data Analytics: Customers looking to run a wide range of near real-time analytics with their choice of Big Data analytics vendors like Hortonworks, Cloudera, Pivotal, IBM and Splunk.
How are the Caches handled in Isilon AFA? Use only the L1 and L2 Caches?
With SmartPools, is it reasonable to assume that the Isilon AFA nodes will be used the most? If that's the case, is there something smart to protect the flash arrays from worn out quickly?
A customer asked about the following scenario. they have a number of S-, X-, and NL- series nodes in their cluster and they are running out of floor space in their data center. Can they integrate Isilon All-Flash into their existing Isilon cluster and move the data to the new All-Flash nodes and then eliminate the old S- and X- nodes to free up floor space, power, cooling etc?
could you talk about how this is accomplished and how disruptive this is likely to be.
This would be accomplished using SmartPools. In order to join the All-Flash to an existing Isilon Cluster, the existing cluster would need to be upgraded to a new version of OneFS to support the new All-Flash hardware. Once the All-Flash is joined to the existing cluster, a SmartPool policy can be used to move data from the S-, X- and NL- nodes.
The Isilon AFA will still use L1 and L2 cache. The L2 cache in particular will be considerably larger because of the large memory footprint of the AFA. Since the node is comprised exclusively of SSD, L3 caching won't apply since all data and metadata will already live on flash.
Where data lives, both the initial write and longer term retention, is based on your SmartPools policies. As mentioned above the larger cache footprint allows us to even more efficiently coalesce writes. On top of that, improved drive firmware and larger SSD sizes means that wear is much less of an issue that it was in the past.
A customer asked about adding additional nodes or chassis to a All Flash cluster.
To start a cluster, one chassis of 4 nodes is required. At that point, to increase performance or capacity does the customer have to purchase another full chassis of 4 nodes?
Or, can they purchase a chassis that only contains 2 nodes?
Starting at 4 nodes for initial cluster, after which increment in "couplets" (2 nodes at a time)
Interested in 'Unmatched efficiency" with 80% and 30% numbers. As listed, with one chassis of Isilon AFA, it can reach 924 TB raw capacity. The 80% storage utilization is about 740 TB. With SmartDedupe, the storage requirements could reach 960 TB (30% increase). Do these numbers make sense?
How does SmartDedupe work in AFA? Is it done via inline process? With inline compression, it should have more effective storage capacity available. Are there some test numbers for it? It would be useful in the stage of storage planning.
This Ask the Expert event has officially ended, but don't let that retract you from asking more questions. At this point our SMEs, after addressing all questions previously posted, are still welcomed to answer and continue the discussion though not required. Here is where we ask our community members to chime in and assist other users if they're able to come up with accurate information. Additonally, you're welcome to post other questions unrelated to the topic of this ATE on the Isilon Product Community.
Many thanks to our SMEs who selflessly made themselves available to answer questions. We also appreciate our community members for taking part of the discussion and ask so many interesting questions.
ATE events are made for your benefit as members of ECN. If you’re interested in pitching a topic or Subject Matter Experts we would be interested in hearing it. To learn more on what it takes to start an event please visit our Ask the Expert Program Space on ECN.
Interested in 'Unmatched efficiency" with 80% and 30% numbers. As listed, with one chassis of Isilon AFA, it can reach 924 TB raw capacity. The 80% storage utilization is about 740 TB. With SmartDedupe, the storage requirements could reach 960 TB (30% increase). Do these numbers make sense?
How does SmartDedupe work in AFA? Is it done via inline process? With inline compression, it should have more effective storage capacity available. Are there some test numbers for it? It would be useful in the stage of storage planning.
Your theoretical math is correct. The 80% storage efficiency with erasure coding combined with a ~30% logical saving with dedup. Note the dedup benefit is dependent on workload so this will vary. SmartDedupe has not changed in implementation. Compression is not supported.
We do not have any test numbers to share, but ill follow up on that ask and see what we can do.
RobertoAraujo1
2 Intern
•
718 Posts
0
October 19th, 2016 08:00
Welcome to this Dell EMC Community Network Ask the Expert conversation. Our discussion is now open for interaction.
When formulating your question please consider the following:
Login or register to engage – You must login or, if you’re new, register to DECN in order to post a question, like, bookmark or follow this discussion. https://developer-content.emc.com/login/login.asp.
Be respectful – SMEs are volunteering to assist you and will provide you with information that they have access to.
Provide details and background – Our SMEs will try their best to provide a full answer for your satisfaction, but will appreciate as much information you can provide about your inquiry.
Keep focus and within topic – The expectation is to receive questions about the topic is that is being presented on the title and description of this event. We would appreciate if questions are kept within those boundaries.
Be appreciative – Many tools available for this, likes, user ratings, follow, bookmark and the good old “thank you”.
You have until November 2nd to post questions.
We’re delighted to know you’ve decided to join our debate and hope you receive the answers you’re eager to find. Thanks!
Phil.Lam
3 Apprentice
•
636 Posts
0
October 26th, 2016 13:00
What Front End NICs are available to be used?
geeta_vaghela
8 Posts
0
October 27th, 2016 14:00
10GbE (SFP+) or 40GbE (QSFP+)
geeta_vaghela
8 Posts
0
October 27th, 2016 14:00
Select "Get Pre-Order Information" from here: http://www.emc.com/en-us/storage/isilon/index.htm
This will route you down the path to get the details you need.
geeta_vaghela
8 Posts
0
October 27th, 2016 14:00
The new Isilon All-Flash platform will provide blazing performance to run a wide range of the most demanding NAS workloads. Examples include:
1) Data Lake: Customers looking to consolidate their unstructured data with a data lake that supports a wide range of workloads including a number of applications that require extreme NAS performance.
2) Media & Entertainment: M&E customer target use cases include: 4K streaming of data to drive higher resolution of viewing content; and large-scale render farms that need to process larger datasets and generate content faster.
3) Life Sciences: Target use cases include genomics sequencing and large-scale microscopy environments that need to process enormous datasets in a short time period
4) Financial Services: Use cases will include workloads that need high performance, low latency and maximum throughput such as ticker analysis that requires near real-time analysis of stock market data on very large datasets.
5) Big Data Analytics: Customers looking to run a wide range of near real-time analytics with their choice of Big Data analytics vendors like Hortonworks, Cloudera, Pivotal, IBM and Splunk.
6) Big Data Analytics: Customers looking to run a wide range of near real-time analytics with their choice of Big Data analytics vendors like Hortonworks, Cloudera, Pivotal, IBM and Splunk.
bang11
17 Posts
0
October 27th, 2016 17:00
How are the Caches handled in Isilon AFA? Use only the L1 and L2 Caches?
With SmartPools, is it reasonable to assume that the Isilon AFA nodes will be used the most? If that's the case, is there something smart to protect the flash arrays from worn out quickly?
scott_owens
60 Posts
0
October 28th, 2016 10:00
This would be accomplished using SmartPools. In order to join the All-Flash to an existing Isilon Cluster, the existing cluster would need to be upgraded to a new version of OneFS to support the new All-Flash hardware. Once the All-Flash is joined to the existing cluster, a SmartPool policy can be used to move data from the S-, X- and NL- nodes.
mdkirkpatrick
7 Posts
0
October 28th, 2016 10:00
The Isilon AFA will still use L1 and L2 cache. The L2 cache in particular will be considerably larger because of the large memory footprint of the AFA. Since the node is comprised exclusively of SSD, L3 caching won't apply since all data and metadata will already live on flash.
Where data lives, both the initial write and longer term retention, is based on your SmartPools policies. As mentioned above the larger cache footprint allows us to even more efficiently coalesce writes. On top of that, improved drive firmware and larger SSD sizes means that wear is much less of an issue that it was in the past.
geeta_vaghela
8 Posts
0
October 30th, 2016 23:00
14. Re: Ask the Expert: Accelerate the Data Lake with Isilon All-Flash
sri_seshadri Oct 29, 2016 4:31 AM (in response to Roberto Araujo)
A customer wanted to know if they could mix drives of different capacities in the same chassis.
In other words, can they have the 15TB drives in 2 nodes of the chassis and the 3.4TB drives in the other 2 nodes of the same chassis?
what combinations of mixed capacities can they have at a cluster level and at a chassis level?
Chassis must be homogeneous - same capacity for both couplet pairs.
geeta_vaghela
8 Posts
0
October 30th, 2016 23:00
sri_seshadri Oct 28, 2016 6:31 PM (in response to Roberto Araujo)
A customer asked about adding additional nodes or chassis to a All Flash cluster.
To start a cluster, one chassis of 4 nodes is required. At that point, to increase performance or capacity does the customer have to purchase another full chassis of 4 nodes?
Or, can they purchase a chassis that only contains 2 nodes?
Starting at 4 nodes for initial cluster, after which increment in "couplets" (2 nodes at a time)
bang11
17 Posts
0
November 1st, 2016 15:00
Interested in 'Unmatched efficiency" with 80% and 30% numbers. As listed, with one chassis of Isilon AFA, it can reach 924 TB raw capacity. The 80% storage utilization is about 740 TB. With SmartDedupe, the storage requirements could reach 960 TB (30% increase). Do these numbers make sense?
How does SmartDedupe work in AFA? Is it done via inline process? With inline compression, it should have more effective storage capacity available. Are there some test numbers for it? It would be useful in the stage of storage planning.
RobertoAraujo1
2 Intern
•
718 Posts
0
November 4th, 2016 14:00
This Ask the Expert event has officially ended, but don't let that retract you from asking more questions. At this point our SMEs, after addressing all questions previously posted, are still welcomed to answer and continue the discussion though not required. Here is where we ask our community members to chime in and assist other users if they're able to come up with accurate information. Additonally, you're welcome to post other questions unrelated to the topic of this ATE on the Isilon Product Community.
Many thanks to our SMEs who selflessly made themselves available to answer questions. We also appreciate our community members for taking part of the discussion and ask so many interesting questions.
ATE events are made for your benefit as members of ECN. If you’re interested in pitching a topic or Subject Matter Experts we would be interested in hearing it. To learn more on what it takes to start an event please visit our Ask the Expert Program Space on ECN.
Ask The Expert – Isilon’s New Releases: IsilonSD Edge, OneFS.NEXT and CloudPools
Ask the Expert: Are you ready to manage deep archiving workloads with Isilon’s HD400 node and OneFS 7.2.0? Find out more about the Data Lake Foundation products
Ask the Experts: EMC Isilon technical content and documentation
geeta_vaghela
8 Posts
0
November 4th, 2016 15:00
Interested in 'Unmatched efficiency" with 80% and 30% numbers. As listed, with one chassis of Isilon AFA, it can reach 924 TB raw capacity. The 80% storage utilization is about 740 TB. With SmartDedupe, the storage requirements could reach 960 TB (30% increase). Do these numbers make sense?
How does SmartDedupe work in AFA? Is it done via inline process? With inline compression, it should have more effective storage capacity available. Are there some test numbers for it? It would be useful in the stage of storage planning.
Your theoretical math is correct. The 80% storage efficiency with erasure coding combined with a ~30% logical saving with dedup. Note the dedup benefit is dependent on workload so this will vary. SmartDedupe has not changed in implementation. Compression is not supported.
We do not have any test numbers to share, but ill follow up on that ask and see what we can do.
Peter_Sero
4 Operator
•
1.2K Posts
0
December 4th, 2016 02:00
> This Ask the Expert event has officially ended, but don't let that retract you from asking more questions.
What is used, SAS-attached SSDs or NVMe flash?
Or has that not yet been disclosed?
-- Peter
Peter_Sero
4 Operator
•
1.2K Posts
0
December 5th, 2016 21:00
Thank you!