3RP Tech Blog

Finally having our OpenStack Private Elastic Cloud event.  Check it out...

We are going to host an event in late June.  What do you think? email or call me.

Encouraging Bad Behavior

XtremeSF (cards) is cool and fast. It was on the slides so it must be true. Scales to 2.2TB and can be partitioned to have a Read Cache on one side and the other a R/W drive letter. Can get real world improvement of 250-450% in latency and IOPS XtremeSW (software) will eventually supper 3rd party cards (FusionIO, Micron, others) XtremeIO - Code 1.05 available in directed availalblity. Missing raid 6, snapshots, or HA Full release in Q4 2.2 GA will have non-disruptive upgrades, HA, VPlex, bCenter, PowerPath, everything. Wicked. 2014 - 100TB bricks and RecoverPoint. Because all replication technologies cannot keep up with performance of XtremeIO, lots of tuning will be engineers over the year Blowout performance for test/Dev database farms where rapid clones are deduced on the fly. Full copies take up zero space. Big Deal for cushing Oracle and SQL environments.

EMCWorld Joe Tucci keynote. Ouch... But true

Forgot to mention on upcoming Isilon OneFS 7.1 release Post process 8k block dedupe. (take that NetApp marketing department) Hadoop Support and others. Link has great overview of features. Will require cluster reboot :-(

SE Breakout Wed for only EMC and partner SE's, No Customers Class #1 MS HyperV 3.0 on Server 2013 Quote "Microsoft may always be late to the party, but they always bring the keg" These are the presentations we have been looking for..... :-) (David @ 3RP is digging this)

(I think marketing will need to punch it up a bit))

Had a 1-on-1 conversation with Pivotal. 3RP loves the PivotalHD hadoop 2.x distro with HAWQ and we now know how the GEMfire and SQLfire blend to the story. Hard to do over the phone.

EMCWorld Experience Center. Anyone want a cheesy t-shirt? At least you got a free beer out if it.

SwiftTest booth was cool. Load testing for SMB2/3 and NFS 2,3,4, and 4.1, load balances, etc Turn your infrastructure to 11 and see when it breaks before you role it into production. Go Daddy uses it before rolling everything before rolling to prod or evaluating vendors. Really cool. Check it out and let me know your thoughts.

XtremeIO presentations (weak). Had great 20min 1-on-1 conversation with an actual customer running XremeIO SF (PCIe card) and said you can get 9GB/sec with out of box support (no tuning). He compared against FusionIO (I am a shareholder) and it was about the same on speeds but the software and failure support on XtremeSF cards blows it away. Great discussions

We met Cody, VPlex Product manager at the EMC experience center. We got the rundown in installation, gotcha's, etc.. and think this product is under appreciated by customers on how to make a 100% uptime data center with manageable costs and labor. Better than we thought....

Last day at EMCWorld. Had great dinner last night with some of our customers. Heading to SE training at the Mirage today to get the meet of the future, no the sales fluff. Exciting!!!!

Multi-Data Point verification is the new trend Airline flight to hotel check in to cab ride says Scott is in Vegas and let him take out money from the ATM. RSA..... Scary concepts and most have no idea.

Pivotal Data Fabric - Greenplum SQLfire Gemfire PivotalHD GemFire - draw ram from many different servers up to 100's of nodes - change I/O problems ($$$) to CPU problems (cheap) Continuous Query (Gemfire) - you can subscribe to it for updates on events GemFire - objects SQLfire - relational HAWQ - Hadoop with Query Greenplum GPFX - connector to Hadoop ILM issues solved via Columnar for cold data and row oriented for hot data on partional level Compression - GP is the data processing horsepower for in-data algorithms (MadLib) - all open source GEMfire => SQLfire => Greenplum => PivotalHD. No problem to fast or too BIG Gemfire is streaming sensors then batch dump to Greenplum Gemfire gets data in, SQLfire gets data out GP extracts subset (microcubes) into SQLfire for realtime GEMfire and work sync over unreliable WAN links --- or so the nerds say

VMware making Big Data Dance Current customer sample running hadoop (40% attendees) Stage 1 - Pilot (65%) Stage 2 - production 1 dept / hive/pig. (35%) Stage 3 - cloud Analytics platform - integrated bi tools, many departments Project Serengetii Deploy Hadoop in VMware - minutes Enter how many nodes and what distribution and how big of resources Proactive monitoring with vCops Challenges of Hadoop in Enterprise 1) Production 2) Test 3) Experimentation What do you do with a noisy neighbor? QOS in VMware PivotalHD is the first Hadoop distro to support Horzontal scaling (qty) and vertical (size) in VM - cool