ORIGINALLY POSTED 4th February 2009
8,501 views on developerworks
If SVC put out a press release every time we qualified a new controller behind the cluster we’d be putting out one every couple of weeks. To claim in such a way that this was a new an interesting thing to do with a TMS RAMSAN is also a stretch…
I am of course referring to yesterdays release from NetApp – which left a lot of the storage folks on Twitter wondering if we’d missed something.
A few others have commented on this in the blogosphere, Zilla has tackled this in his unique way, and you can tell from the tone that Martin is seriously dumbstruck.
Looking at some of Martin’s points made me go an check how much I could discuss here about things that are mid-way through SVC qualification. One of which is, guess what the RamSan-500. Supporting a storage controller behind SVC comes under two categories. Internally we refer to these as the ‘tier1’ and ‘tier2’ vendors. Nothing to do with the storage devices they produce, but the vendor themselves. So IBM, EMC, HDS, HP, Sun and NetApp are all in the ‘tier1’ category – the main names in the storage industry, so we try and ensure we have at least GA+90 support when they release new products. We have a good relationship with “most” of these companies, and they see the benefit to their existing and new customer bases to being supported behind SVC. The second category of vendors are the smaller brands, not necessarily less important, but they are less prevalent in the market place, and so in the SVC bids come up less often.
As with most vendors we have an RPQ process that allows customers, business partners and account teams to submit a request for support of something not on our matrix today. There is a finite amount of test resource available for these tests, so a priority scheme is used, at its simplest based on the number of requests – or the value of the bid etc etc. Once something has been agreed by the Interop teams it will enter test. Several months ago we took delivery of a RamSan-500 from TMS to qualify behind SVC as we saw an increasing number of requests for said support. I wouldn’t even be mentioning this, but given that we are more than halfway through test and no major issues have been seen yet – in light of yesterdays NetApp announcement, and Martin’s comments, I thought it was worth mentioning.
I stated above we don’t make big splashes about new controller support, and it certainly never crossed our minds to try and make people think the qualification of the TMS product behind SVC would be a major thing, to us its another controller, a pretty fast one, but certainly nothing earth shattering.
An organic solution would be much more interesting…
Leave a Reply