tag:blogger.com,1999:blog-2987385736576623269.comments2022-12-03T18:17:31.695-06:00Retooling the DatacenterUrbanhttp://www.blogger.com/profile/06019262251107123583noreply@blogger.comBlogger21125tag:blogger.com,1999:blog-2987385736576623269.post-18742057582102061292013-07-10T15:44:42.533-05:002013-07-10T15:44:42.533-05:00I'm thankful that i have been searching for th...I'm thankful that i have been searching for this information and finally i got it. Thanks for the information.<br /><br /><a href="http://byodsecurity.org/" rel="nofollow">Security</a>Jahidul Islam Bhuiyanhttps://www.blogger.com/profile/00334753899980382448noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-82078154169922876552012-06-17T20:06:33.321-05:002012-06-17T20:06:33.321-05:00Mac,
I couldn't agree more, that's why we...Mac,<br /><br />I couldn't agree more, that's why we separate disaster recovery from business continuity. BC is Long-Distance vMotion, or Workload Mobility. We still use traditional SRM-type recovery for going beyond the local metro to keep disaster recovery. That is specifically what you're talking about. <br /><br />You can never sacrifice disaster recovery and LDVM doesn't aim to do that.Urbanhttps://www.blogger.com/profile/06019262251107123583noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-71404274450803125432012-06-17T19:59:52.638-05:002012-06-17T19:59:52.638-05:00Xaminmo,
In an enterprise-class box (SVC) FC-IP-b...Xaminmo,<br /><br />In an enterprise-class box (SVC) FC-IP-based replication isn't as big of a deal. Bandwidths are big enough to warrant a faster level of replication.<br /><br />The v7000 is a mid-tier box. In it's space, EMC, NetApp, Dell Compellent, Equallogic, Lefthand all have native-IP based replication. The v7000 has to add on $100,000 of extra gear to give native replication, or follow the backup/restore have of Fastback. I was hoping that we would have seen this by now. I've been holding out on selling this solution purely based on this fact. Over 80% of my customers replicate to a DR site.<br /><br />This is an eloquently designed solution. The auto-tiering would benefit from three tiers, but it's adequate. The GUI is the best on the market. Adding in IP-based replication would make this one of the best solutions on the market. I would have thought that iSCSI, adding the IP-protocol stack with SCSI-3 support would have been enough to enable replication. RACE would lower bandwidth requirement, which would be a plus.<br /><br />I've been waiting for IBM to add this, then I will start designing solution based on the product. For now, it adds too much cost to justify the box.Urbanhttps://www.blogger.com/profile/06019262251107123583noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-8125726281223105412012-06-17T19:57:51.555-05:002012-06-17T19:57:51.555-05:00This comment has been removed by the author.Urbanhttps://www.blogger.com/profile/06019262251107123583noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-70345813822555545632012-06-15T12:20:04.917-05:002012-06-15T12:20:04.917-05:00IP based replication was dependent upon RACE. Sin...IP based replication was dependent upon RACE. Since realtime compression just made it into 6.4, I would expect it to be 6.5 or 6.6 (or 7.x equivalents) before that came about.<br /><br />And yah, Fastback for IP replication? Unless something magic is going on there, that's not replication. That's a backup repository that still has to be restored somewhere, and still has architectural performance limits. Sure, there's "Instant Restore", but that's just a fancy way of making the backup repository a read-only mirror copy for quicker access to the data. You still have to restore it somewhere, multiplying the storage cost.MaxiPadhttps://www.blogger.com/profile/04650660045425350918noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-49018281900477195432011-10-19T15:54:38.879-05:002011-10-19T15:54:38.879-05:00Reliance on data management systems has its own ad...Reliance on data management systems has its own advantages and disadvantages. It may make our work faster, but there are also challenges in overcoming natural disasters and maintaining computer security. Even data systems such as Long-Distance Motion need back-up systems in case emergency strikes.<br /><br /><br /><br /><a href="http://www.datacenterjournal.com/data-center-management-jobs-more" rel="nofollow">Mac Pherson</a>Mac Phersonhttps://www.blogger.com/profile/07960130410045442181noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-1281757243044080962011-09-19T23:32:49.627-05:002011-09-19T23:32:49.627-05:00Almost a year later and we still have seen mum on ...Almost a year later and we still have seen mum on native IP-based replication. Many people are put out by this.<br /><br />IBM really need to step up and commit to a date for this feature to take this otherwise elegantly designed box seriously.<br /><br />Fastback requires too much overhead, at least two servers and limited OS support.Urbanhttps://www.blogger.com/profile/06019262251107123583noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-8702144717596116942011-07-28T10:16:16.249-05:002011-07-28T10:16:16.249-05:00Thanks for the clarification. I was familiar with ...Thanks for the clarification. I was familiar with the SCOM plugin, but didn't know it could play the tie-breaker role. <br /><br />I'll be sure to revisit my current MetroCluster customers, since they're older than the SCOM plugin.Urbanhttps://www.blogger.com/profile/06019262251107123583noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-71098132813295278922011-07-28T09:35:29.624-05:002011-07-28T09:35:29.624-05:00Disclaimer: I am a NetApp Employee
First of of al...Disclaimer: I am a NetApp Employee<br /><br />First of of all Well done on the Blog.<br /><br />Just wanted to clarify that the NetApp Metrocluster solution can have a Tie-Breaker which can react to a split-brain situation. The tie breaker solution is part of Microsoft SCOM though our On Command Plugin for Microsoft in case the customer is using it or it can be installed on a separate node in case the customer doesn't have SCOM in their environment.Anonymoushttps://www.blogger.com/profile/18253778494401051261noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-80704777544614636952011-07-25T06:23:13.536-05:002011-07-25T06:23:13.536-05:00Thanks for the clear explanation and the excellent...Thanks for the clear explanation and the excellent examples. really good post. keep it up..<br /><br /><a href="http://www.newtonit.co.uk" rel="nofollow">Video conferencing uk</a>Kimhttps://www.blogger.com/profile/17472685254708001583noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-76951066292818677902011-04-09T10:49:20.443-05:002011-04-09T10:49:20.443-05:00Jeramiah,
Thanks for the comment. There are diffe...Jeramiah,<br /><br />Thanks for the comment. There are differences between the architectures of FAST Cache and FlashCache, as there are differences on how reads and writes flow through each system's firmware. I thought I stressed that FAST Cache works with both reads and writes. FlashCache doesn't, but I'm not going into a detailed explanation of how WAFL works. Conceptually they're the same -- an extension of cache. Technically there are many differences.<br /><br />FlashCache needs to warm up, it isn't persistent across reboots, it doesn't cover writes. FAST Cache doesn't need to warm up because it is persistent across reboots (except the very first time), it does handle writes. FlashCache is on the PCI bus is has very very low latency. FAST Cache is on SSD and has SCSI and SAS protocol overhead as well as RAID-1 mirroring.<br /><br />There are differences, but they're both an extension of cache. That was the takeaway.<br /><br />Thanks for your comment. I appreciate it and will check out your blog.<br /><br />UrbanUrbanhttps://www.blogger.com/profile/06019262251107123583noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-62715761703302630082011-04-08T10:31:59.277-05:002011-04-08T10:31:59.277-05:00Urban,
(Disclaimer, I work for VCE on the SP and ...Urban,<br /><br />(Disclaimer, I work for VCE on the SP and Vertical Solutions team)<br /><br />It's good to see the external validation both of the technical side of the VNX/VNXe product launch as well as of the way it was rolled out. It was definitely designed to make an impact on the market! Stay tuned for more on how it plays with the Vblock line as there's some goodness coming there as well.<br /><br />One small comment I'd make is your comparison of FAST Cache to FlashCache from NetApp. PAMII uses a DRAM-on-PCIe model that provides for read-cache only. EMC FAST Cache uses standard SSD drives to extend the array cache for both reads AND writes. Each implementation has its benefits, but there are enough differences that I wouldn't call them direct analogies of one another. If you need more info let me know, I'm trying not to get too far into the (FUD-infested) weeds on it. :-)<br /><br />Hope this helps, thanks again for your feedback!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-63067017434510141002011-01-14T17:23:00.399-06:002011-01-14T17:23:00.399-06:00I'd be interested on your thoughts about any r...I'd be interested on your thoughts about any risks introduced by integrating deduplication into the fabric of an environment, particularly when deployed at various/multiple layers simultaneously as you suggest may be the case.David McClellandhttps://www.blogger.com/profile/05197877784983621206noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-40684176215400304442010-11-09T20:15:57.136-06:002010-11-09T20:15:57.136-06:00Am I the only one who's noticed that if you tu...Am I the only one who's noticed that if you turn the XIV logo upside down, it spells AIX?Urbanhttps://www.blogger.com/profile/06019262251107123583noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-87858654394655735572010-10-25T02:54:38.951-05:002010-10-25T02:54:38.951-05:00Yep! I am also interested what would IBm do on thi...Yep! I am also interested what would IBm do on this. Thanks for sharing. This is really a good heads up for IT people like us. I will also share this to my friends. <br /><br /><a href="http://www.advalem.fr" rel="nofollow">Datacenter</a>Altrics hostinghttps://www.blogger.com/profile/11171171734468262592noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-77232303925785965202010-10-19T09:24:30.125-05:002010-10-19T09:24:30.125-05:00It will be interesting to see what IBM will do bet...It will be interesting to see what IBM will do between the SVC and the V7000. Could the V7000 participate as an iogroup? If the SVC truly is scalable to 32 nodes, we may now have a good case for it.Urbanhttps://www.blogger.com/profile/06019262251107123583noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-46789266455238415942010-10-13T17:20:51.497-05:002010-10-13T17:20:51.497-05:00Hi Urban,
We do support IP replication by means o...Hi Urban,<br /><br />We do support IP replication by means of Tivoli Fastback. This can be ordered as a bundle feature with the Storwize V7000. While this requires a server or guest to install the Fastback software, it does allow incremental replication over IP. The native IP replication is in the roadmap.<br /><br />As for FC replication, yes we have used the entire SVC code stack here, adding in the RAID and drive support, however as this does give us all the interop support of SVC on day1, we do need to make some changes to allow a "controller" to be seen as a replication peer in SVC. This was not deemed as a must have for release 1, but is again in plan.<br /><br />Glad you like the product in general.<br />BarryBarry Whytehttps://www.blogger.com/profile/08003827517858074940noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-55891630757162144842010-09-30T18:53:56.549-05:002010-09-30T18:53:56.549-05:00As a customer I hope that XIV doesn't abandon ...As a customer I hope that XIV doesn't abandon its commodity parts architecture. This is what makes XIV so attractive ... $/TB, stability, availability. All the improvements listed above may be nice as options for the subsets of corporate data that require the performance boost. XIV in it's current state handles general SAN Storage requirements quite nicely.Slackhttps://www.blogger.com/profile/04252202352854827970noreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-70320347579769408392010-09-14T03:21:38.155-05:002010-09-14T03:21:38.155-05:00I would be most surprised if they wouldn't alr...I would be most surprised if they wouldn't already be working on the things you mention. So let's wait and see. Regarding the SSD's you suggest, I don't think it'll be inexpensive anymore then.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-57752169640342730942010-09-02T23:29:03.722-05:002010-09-02T23:29:03.722-05:00Passion and technology are two things I hold much ...Passion and technology are two things I hold much respect. It's nice change of pace -- In fairness to your opinion, Let's just say I don't share the same rose colored glasses. Food for thought - An alternative perspective. If XIV innovated as you designed - improving the data classes and interconnects to allow for varied workloads - MultiTenancy Storage would have a home - This innovation would be called Quality of Service and The Storage Array would be called Pillar. No offense to Moshe but XIV lacks one key ingredient for being a repeat Symetrix/EMC - Erez Ofer - The prodigy software engineer responsible for EMC open systems success. Software being the paramount weakness. Symmetrix was a true innovation of it's day - Unified Storage with MVS and Unix data on a single shared array. Enterprise Storage ---<br /><br />Having worked for EMC my role provided several opportunities to collaborate w/ Erez, Moche, Donatelli etc... All information is not created equal. Why then build storage that services requests FIFO ?<br /><br />Executive platinum air travel is the best analogy for Storage I can offer.<br />The card jumps the queue and from last in line to the First in line. Based on Priority. 120.000 IOPS concurrently served in accordance to their QoS priority - FIFO has no place in SSD, SATA, FC intermixed storage systems. Software is a requirement. <br /><br />Priority of DATA based on the application and dynamic allocation of CPU, CACHE and resources is the future. Moshe will surely induct to the Storage Hall of fame someday. Just not for XIV. <br /><br /> Larry Ellison has spend 10 years and a respectable sum of capital on Pillar. They are Private, Not for sale and growing at an aggressive clip. Quiet Giant that walks softly and yields a Big Stick! <br /><br />When you've had XIV for breakfast - It's a game changing meal.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2987385736576623269.post-17225022385248771052010-08-27T07:50:03.954-05:002010-08-27T07:50:03.954-05:00I'd be happy if they offered a parity disk per...I'd be happy if they offered a parity disk per tray, or a 3rd mirror position for business critical data.<br /><br />The non 50% 50% distribution between interface and data nodes is a little bizarre too... would like to see it symmetric.techmutehttps://www.blogger.com/profile/02613472355641438692noreply@blogger.com