Elisabeth Stahl on Benchmarking and IT Optimization

Oracle’s Storage Riot

with 2 comments

Last weekend there was a flash mob in my local favorite shopping district. A few weeks before that someone got mugged outside of my favorite Turkish restaurant. And last night, while waiting for our town’s amazing fireworks display, a huge fight started that almost started a stampede. And I don’t live in Bed-Stuy.

It was decided that the solution would be a restrictive curfew for all kids under 18. Good idea, right? Except when the curfew is 6PM — which you can imagine has caused an uproar.

Which has gotten me thinking — is anywhere really safe anymore?

Which is how I used to feel about storage. Unlike their server brethren, storage was pretty straightforward. You bought it, you attached it, and you moved on. Occasionally there was a Storage Performance Council benchmark to help guide your way.

And then just like rowdy teens using Twitter, things like big data and petabytes appeared.

Last week Oracle held a Storage Strategy Update. In it Mark Hurd says he likes numbers. He also says “we think we can run” applications up to 10x faster at 1/10 the storage capacity. 10 is a nice number. Does claiming a “big number” or a “step function” make it any more real? Note that Mark says it won’t work for every application. Which ones does it actually work for?

Where did the performance claim come from? It looks familiar – I’m thinking recycled Exadata claim with no data. And speaking of Exadata, it’s also interesting that in Oracle’s strategy presentation the number of systems SOLD were highlighted for each member of the storage portfolio – all except Exadata where only INSTALLS were mentioned !

IBM continues to be #1 in the Storage Performance Council SPC-2 benchmark(1) which demonstrates the performance of a storage subsystem during the execution of business critical applications that require the large-scale, sequential movement of data.

Oracle talks exabytes. IBM is really ready for the zettabyte age. Now that’s “extreme performance.”


(1)The SPC-2 IBM System Storage DS8800 result of 9,705.74 SPC-2 MBPS and a price/performance of $270.38 at a ASU Capacity of 71,536.975 GB in a RAID5 configuration consisting of 2 – 4-way SMP processing clusters 128 GB memory/cache per cluster (256 GB total) 16 – 8 Gb FCP/FICON Adapters 64 – 8 Gbps FC host connections, 4 per FCP/FICON Adapter, (16 connections used, 1 per adapter) 16 – FC/AL Device Adapters (8 pairs) 64 – 8 Gbps FC backend connections, 4 per FC/AL Device Adapter (64 connections used) and 768 – 15K RPM 146 GB disk drives (2 DS8800 expansion units). Source: Result current as of 7/5/11.

The postings on this site solely reflect the personal views of the author and do not necessarily represent the views, positions, strategies or opinions of IBM or IBM management.

technorati tags: , , , , ,,,,,,,,,,,,,,


Written by benchmarkingblog

July 5, 2011 at 5:04 pm

Posted in Exadata, Oracle, storage

Tagged with , ,

2 Responses

Subscribe to comments with RSS.

  1. “…storage was pretty straightforward. You bought it, you attached it, and you moved on.”
    – Spoken like a true server chick! For the last 10 years I’ve felt that about servers. Storage is wher ethe action is : )


    July 5, 2011 at 5:55 pm

  2. We have acknowledged your previous benchmark and responded accordingly. We have beaten your DS8800 benchmark for 7x less cost…

    Also if you truly understood HCC and what it does to oracle queries then you would be less questioning how we get 10x faster. Hybrid Columnar Compression is very real…

    Darius Zanganeh

    April 17, 2012 at 6:47 pm

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: