Discussions ranging from enterprise storage, data-centre infrastructure and cloud infrastructure to TVRs and motorsport in general.
Expect general grumpiness, frequent rants, and plenty of complaints & challenges re vendor FUD and hype. Frequently heard shouting "show me the TCO & ROI"...
In the late 1990s a crack storage admin unit was sent to a data-centre
by senior mngt for a issues they didn't create. These men promptly
escaped from a maximum security site to the UK underground.
Today, still wanted by the vendors, they survive as admins of fortune.
If you have a storage problem - if no one else can help - and if you
can find them - maybe you can contact : The (storage)A-Team.
If you're in the UK, a storage customer, and want to meet like minded storage admin peeps then sign the NDA, request to join and get ready to contribute on a 2-3 times per year basis...
If you're a vendor or reseller then you certainly can't do sales / marketing pitches - but, if you're lucky, maybe you can engage with the team to review your product, give feedback and sort your RFE & roadmaps...
[Disclaimer - the forum used to have facilities & logistics funding (repairs to the team van, all the cigars & gold chains costs you know!) and assistance from EMC (I'm currently unsure) - but certainly no sales pitch or content control]
[Double disclaimer - my terrible time & diary mngt has meant I've been a very underground figure for too long in this forum]
Procurement & purchasing models? (eg large pre-provisioned boxes, small boxes with chunk growth, ad-hoc project based etc)
Do you envisage this situation changing in light of technologies such as :-
Thin provisioning
Wide striping
Automated lun/sub-lun tiering (eg FAST, TSM etc)
Space reclomation (eg ZPR etc)
VM/FS improvements
Some form of improved & useful SRM tools (yes I realise this is a big wish)
I'm interested to know, as we have historically tried to adopt standard RAID layouts, device sizing and LUN sizing - with the view that this eases, simplifies & speeds up storage operational and diagnostics tasks. But increasingly I'm thinking that this matters far less with the storage technologies now available.
Ok so some people will have seen my various twitter posts whilst attending SNW-E in Frankfurt a couple of weeks ago, so this is a bit more of a set of feedback / thoughts.
Firstly some housekeeping - I posted on twitter on Tuesday that I questioned the conference attendance statistics (http://twitter.com/ianhf/status/5227700819)- now to my very pleasant surprise a few senior people have taken note to this. My comment was not intended to be questioning the validity of the statistics, but rather 'how it felt being an attendee'. I'd discussed this with a few other people and the common feeling was that it appeared to be less busy than previous years, my suspicion is that this feeling was generated by the revised format (more on that later) and thus not getting visibility of all the people at once.
Right so now to some thoughts :-
1) This is this 4th or 5th year I've attended - so I like and value the conference greatly
2) I've always preferred this conference over UK & USA based shows - the 'session & education first' format works well for what I want (not interested in vendor sales pitches)
3) It would be a conference I'd feel ok paying some form of registration fee (and that's very rare for me)
4) Some rough general feedback thoughts are :-
I felt I really missed the 'keynote' opening and kick-off sessions, whilst this gave a more relaxed and informal feel to the event, it also gave an impression of being a smaller slightly less organised event
The Tuesday morning schedule felt 'light' of interesting topics and yet the afternoon schedule had many overlaps of conflicting topics
The hands-on lab sessions weren't obvious to me, it was only on the Wed that I realised there were any at all this year
I really wish conferences would declare if there are going to give out a rucksack or bag in the booking materials - too often I either have no bag or many bags
Free WiFi in the building and conference session rooms was yet again lacking, and the WiFi that was in some places appeared to be astronomical in cost (€49 was the cost I saw) - needless to say my 3G got a hammering
I fully realise this is a European conference, but it does always surprise me re the number of international technology companies that only bring German product literature - surely it's not that hard to bring a qty of Intl English materials? (given that English works as the common language for other European nationals)
Thankfully the conference has resisted the current conference fad of de-focusing and becoming general IT infrastructure events with width but little depth - it's good to have events that still think about depth but equally don't ignore the width (eg this year's cloud infrastructure theme worked well)
It's always good to see @stevedupe @stephenodonnell in action, the ESG style always works well for me, and certainly resonates well with my experiences
The lunch on Wed seemed to vanish very quickly - meaning I missed out :)
In future I think it would be useful to provide :-
A memory stick or DVD with copies of all the slides and all the exhibitors marketing materials on - a single information bundle - so much easier to annotate slides on the fly
Of course a blogger's lounge would naturally be appreciated :)
The positive is that as usual the sessions were good, with the SNIA sessions providing good time the think over topics - even if a lot of them weren't stretching knowledge, it's always good to spend time refreshing and concentrating on topics without the phone ringing :)
The biggest problem I had was not having enough time within the tighter 2 day schedule (as opposed to previous 2.5/3 day) to meet with the various peers and industry experts & speakers - I need to find a better way for next year.
Overall whilst I think the new format improves a number of areas, something in it didn't click for me, but all being well I'll be there again next year :)
"PowerLink -> Home > Support > Interoperability and Product Lifecycle Information > Interoperability Matrices"
HDS http://www.hds.com/corporate/resources.html
·API / XML driven interface for licenses :-
a) To be able to remotely programmatically determine each & every licensed featured installed / not installed within the device
b) The be able to remotely programmatically determine the license metric and it’s usage for each & every licensed featured installed / not installed within the device
·API / XML driven interface for environmental impact
a) To be able to remotely programmatically determine the real time energy consumption by the asset (power etc)
b) The be able to remotely programmatically determine the high, low & average energy consumption by the asset (power etc) over a given period
IBM
·Would be very interesting to use the SVC to perform Raid5/6 over multiple arrays - to remove the array enclosure as an SPOF in a data-centre
Microsoft
·SQL Server formally & fully supporting NAS for it's database storage
·Formally & fully supporting 3rd party backup/recovery tools for it's database products
Zimbra
·Support of object storage (Castor, S3, Atmos etc)
·zBackup to operate multi-threaded and support for backup / restore of 10TB database messagestores
Caringo
·Full support for running Castor as a VMWare guest image
·Castor AWS S3 compatible API option
·Castor support for Zimbra, Sharepoint & Exchange+Enterprise Vault
·TCG SSC Oasis support
EMC
·Centera usability & readability of config reports
·Ability for SRDF sync replication with delayed updated at remote site
·Atmos support for Zimbra, Sharepoint & Exchange+Enterprise Vault
·Atmos support for an AWS S3 compatible API option
·Networker full support of Atmos as target
·TCG SSC Oasis support
·Greater qty of Ethernet ports on NAS devices
·EMC Powerlink to present at least the same deployment statistics and uptime information as NetApp NOW site
a) Enhancing the current information presented at :-
"PowerLink -> Home > Support > Interoperability and Product Lifecycle Information > Storage Target Revisions and Adoption Rates"
b) To include by product, by firmware release :-
Release Date
Number of customer systems currently running this release
Number of customer sites currently running this release
Average run-time days per system
Qty of issues, by severity
c) To make the following additional information also available via an XML interface :-
Product major & minor (eg DMX & 4), Code Name, Code Rev, Release Date, # or % product running this release, # or % sites running this release, Average up-time, Target / Recommended Release
HDS
·To have an equivalent of EMC PowerLink or NetApp Now for self-support and information access
·API reporting of hot-spare qtys correctly on USP range
·API alert reporting for service processor utilisation on AMS range
·Ability for sync replication with delayed updated at remote site
·TCG SSC Oasis support
·Support for Thin-Provisioing in USP-1100
NetApp
·TCG SSC Oasis support
·Release the software only variant of OnTap as a commercial product (capacity limited if needs be)
·Greater qty of Ethernet ports on NAS devices
·Much larger aggregate & flexvol capacity sizes without having to upgrade to v8
·Support for OnTap v7.3 & v8 on older generation equipment and also for newer products (eg 2020 & 2050)
·Support for native OnTap 'in box' tiering of a file-system over multiple disk types (eg to support FAN)
Cisco
·'terminal release' concept for SANos (check current name) - where support partners must converge upon a certainly release variant within X months (to aid interop)
Twidroid
·To have a "eMail tweet" (inc URL to tweet) that integrates with your google mail account
·To have a "Reply All" option against each tweet
·TwidroidPro posts to be 'sent from' TwidroidPro rather than just 'Twidroid'
·To correctly support Androids 'select & hold' ability to correct default spell-check suggestions
TweetDeck
·eMail tweet to include URL to tweet
·Option to save state of currently 'in memory' tweets upon shutdown, and reload upon restart (I'd pay for this feature alone)
·UK spellchecker
·Ability to set frequency of refresh of searches (in similar fashion to DMs, Mentions etc)
Snarfer
·An update to code?
·Better support for database store to prevent / fix corruption
In this brief blog I'd like to discuss the topic of product RFEs (Requests For Enhancements) - now as far as I'm concerned this is very much one of the most ignored and abused topic areas of the customer / supplier relationship.
Some tiny examples :-
One supplier I know took 4 years to process an RFE that they admitted was both useful and easy to do, but somehow it never quite made the development release train - and nobody could explain why.
Another recently updated their internal RFE system, and in the process simple deleted the details of all previous RFE requests
With one supplier I have at least 3 different ways that I'm required to register RFEs - as the method is different for each product (but feels like it's totally random)
Frankly the status of RFE management staggers me. In @SteveTodd's recent book Innovate with Influence he quite rightly points out that a lot of innovation comes from customer demand - and yet for many vendors the method for accessing and capturing the customer requests is opaque, fragmented and often best described as broken.
Suppliers rarely seem to demonstrate they include consideration for a product's existing customer RFE lists when designing next generation products - this may occur but the communication & dialogue with the customer is sporadic and often after the next product has been 'feature locked'.
The process seems to be normally 'handled' by the local technical representative on the account, resulting in a lot of human processes & resource consumption, interpretations, delays (always dropping down the priority list), difficulties for global orgs (customers & vendors) to align on RFE priorities etc. Thus sorting this out will not only help get (some of) the right features progressed in the right timeframes, but should also save both parties money due to less people effort associated with the 'process'.
Some simple things here could make a big difference, for instance :-
Operating the RFE handling under a clear process and SLA
Publishing a document explaining the process and SLA
Using a standard electronic template for information capture
Having a common format for agreeing & valuing the benefit and priority to both customer and supplier
Understanding that it's key to have regular communication and update on the status of the RFE
Committing to a finite time-line for a decision (ie yes now, yes future, possible, private funding, never etc)
Of course, exactly the same comments and disappointments exist for most vendor's Interop & RFQ (Request For Qualification) handling as well - some key additional points that would help here are :-
Up-front published & consistent clarity over the information required when raising an RFQ
Clarity over the terms & definitions of an acceptance or rejection
Maintaining a customer accessible database of their existing RFQs so that they can be tracked over time and checked for ongoing validity, updates, being superseded, and included within support contracts
Putting RFQs into the standard interop documents ASAP after qualification
Providing a customer accessible and downloadable version of the interop matrix/documents - including maintaining customer access to historical versions (as a large number of RFQs are with older tech)
Providing a greater level of detail within the interop documents for those products that don't have a 100% fully supported status (ie tested but failed, not tested, worked but no longer commercial support etc)
It also continues to surprise me that at vendor's customer council sessions they rarely, if ever, discuss the topic of RFEs - both process and/or content. I'd have thought these would be an ideal forum for suggestions, discussion, refinement, prioritising & voting on RFEs - I know one user forum where this happened and in effect the 30 customer companies jointly agreed on a 'prioritised top 10' RFE list for that vendor and issued it as a single entity. Surprisingly enough that list got focused work on and all the entries were completed within 6 months - a win-win for all!
The topic of more public and peer reviewable RFE lists certainly intrigues me - again most vendors have online web portals of some description for their support forums, yet rarely do they allow customers to view, comment & contribute on RFEs. The exception appears to be some of the newer/smaller companies, who make use of web technologies to perform this. One such technology is http://uservoice.com/ which allows customers to submit, comment/refine, and vote for RFEs - a perfect example being the RFE list for tweetdeck at http://tweetdeck.uservoice.com/pages/1192-general Once again my question is why major vendors don't make use of such simple and positive customer engagement methods?
Yes there is a danger of 'shiny bauble' syndrome, or indeed 'vendor bashing' with RFEs - but if managed and facilitated well, with the right people, then this really is an area that massive strides can be made in short periods of time.
Yes of course there are some interesting areas re IPR ownership, but they tend to be handle by the inter-company framework agreements, or for smaller relationships the terms & conditions etc.
Naturally this is also regularly overlooked in the TCO/ROI cost and benefit calculations - and anybody that's had RFEs eaten by the 'chaos demons' or RPQs impacted by the 'time-delay elves' would certainly agree that there are costs & benefits to be measured here!
Now show me a supplier that openly promotes and communicates a single RFE (& RPQ) process, with user self-serve & transparent status checking and that operates under an SLA - and I'll be a very happy man waving purchase orders!
Ok so those who followed Oracle's Exadata v2 pitch will have seen something mentioned in the specs re FlashCache and it's use of SSD storage. There's little major stuff around but some brief info can be found at :-
Now the part that is of particular interest to me is, if or when, this feature is made available as part of the normal Oracle 11gR2 release train.
This slide from http://www.flickr.com/photos/fenng/3883028359/in/photostream/ appears to show that the FlashCache is interfaced through a file location, thus one assumes it could be stored on any device that is file-interface accessible (eg an array based SSD or a host based SSD).
Now this is where it gets interesting for me, in that in a similar way to OpenZFS L2Arc , a local server SSD could be used as a high speed L2 SGA extension without putting any data at risk.
Server SSDs are supposedly not in the same league as array EFDs, but ignoring that (be it real or not), the price point of SSD in servers (compared to that of those in arrays) combined with this very easy use case model (both setup and ongoing mngt very simple) and lack of risk of data loss (re these use cases), certainly makes it a very intriguing proposition and something on my 'benchmark' and TCO evaluation list very shortly.
My test case would be to take an x86 server running OpenSolaris and place 2x256GB SSDs into it and then run Oracle Swingbench & Orion with :- no SSD usage, 1 for L2Arc, 2 for L2Arc, 1 for FlashCache, 2 for FlashCache, 1 for L2Arc & 1 for FlashCache. Also interested to test 2xFlashCache with ASM Vs 2xL2Arc without ASM. Intrigued to see what the performance impacts will be :)
Naturally I'm also keen to understand & compare these costs & benefits against those of Oracle InMemoryDatabase (eg TimesTen) and array based SSD.
My personal suspicion is that using the SSDs for L2Arc & FlashCache (if / when available) will deliver good benefits cases, and then once the storage array firmware has achieved it's required SSD mngt & usability maturity we'll be able to use array based SSDs in addition to server based for even further combined benefits.
I'd be interested to hear about other people's experiences, thoughts & plans in this area...
Well those who know me may well be staggered by the comments that follow, so get a cuppa, sit down, brace yourself and let's begin :-
I might actually have started to think there is a value case for SSD drives!
Now, this is not in storage arrays - certainly not until array software comes along that can transparently and automatically use these, and when so called 'EFDs' become financially viable (rather than 20x cost of equivalent FC capacity). I'm expecting this to be mid 2010...
But, the value case I've now had first had usage & benefit of is in personal laptop computers.
So earlier this week I had the pleasure of my office IT support dept using HP Radia to roll-out office2003 patches on-top of my office2007 install. Thus rapidly borking my HP tablet PC to a level where it is still being reformatted and rebuilt as I type (3 days later).
Now the interesting bit is that they gave me a loan laptop - a Dell (yes I swore a lot), that despite having less RAM and slower CPU, this laptop appeared to operate in the real-world much faster.
A quick peek and I discovered the Dell was actually using a 64GB Samsung SSD, which if nothing else has gone to show what a poorly written app MS-Outlook2007 is - with the end-user client application being almost entirely IO bound.
The same dataset (mix of OST & PST of about 30GB all offline local) takes about 15mins to load on normal drive, about 1min on SSD. Naturally this is a rather specific use case being a single spindle, mix of multiple IO etc, large % of disk capacity utilised etc.
Even so this has changed my laptop usage feeling so much I'm off to buy my own SSDs (out of my own pocket) to put into my work laptop ad my home gaming rig.
So who has any recommendations for 128GB SSDs for HP 2710p Tablet (work) and normal SATA interface (home)? I've been noodling at these :-
Scottish pessimist - IT infrastructure architecture and strategy in large telco.
Storage is my speciality but am also a stealth Rhubarb picker, Cricket watcher, TVR driver and Laser sailer.
The views & comments on this blog are my personal ones, and are not vetted or represent any past, current or future employers. Oh and the speiling and gammar could be a bit duff :)