The interview went very well. There were questions about tape, where we discussed the advantages of IBM’s LFTS technology for more advanced tape usage, we discussed the direction data deduplication will go as well. Noska’s view was that there hadn’t been any advancement in data deduplication in the last 5 years. I told him that for secondary storage, backup, that he is right, I also told him that the real advancement to deduplication will come when it is ready for primary storage. Today deduplication isn’t ready for primary, but it will be soon.
On Monday the 13th we traveled to visit Avnet. They are a great IBM partner. Like most partners they have a very large SMB install base and also like a lot of SMB feedback I have been getting, they are looking for a building block solution that has all of the software features implemented as a part of the stack. SMB and Enterprise alike are starting to realize that the value in any array is becoming the software stack that makes the hardware, efficient, optimized, flexible, and dynamic. IT’s job continues to get more and more challenging with developing strategic initiatives for the business to make them more competitive and it is the job of the vendor to make sure these solutions are as optimized and cost effective as possible.
We also visited DHL. These guys have one of the greatest datacenters I have ever visited. They are very advanced and push a lot of data. The do some very strategic logistics for a number of companies in Europe and Asia. They, like many others have a number of challenges. Since my blog post about “The 5 Most Interesting things at VMworld” (#4) I heard something very interesting today. I asked “What is your most challenging storage issue”? He told me that storage was not is “most difficult” challenge. Storage efficiency was important to him in order to keep driving down costs for his organization as they deliver a service to the different groups that make up DHL, but his most difficult challenge was with server I/O in his VMware environment. If you read #4 in my post, regarding Proximal Data, this is exactly the issue the address. As VM instances grow on the physical servers, the I/O starts to become the big problem. DHL runs over 4000 instances of VMware and as the business demands more applications and application resources, they are bound by the I/O of the server, which also causes them to WAY over provision their storage for performance reasons. This is very time consuming, management intensive and expensive. The combination of a solution like Proximal Data as well as compression can help them optimize their infrastructure to save money and deliver better, more cost effective services to their lines of business.
On the lighter side, I spend the weekend in Prague. What an amazing city. The weather was fantastic and I was able to take a lot of great photos. I walked around Prague Castle, ate some authentic Czech food, visited the memorial for the Czech hockey players that passed in the Russian plane crash and met some pretty interesting people. You can check out some of my photos of Prague at www.facebook.com/skenniston. Coincidentally the photo above shows the "Golden Lane" where the Alchemists worked to turn anything they could find into gold in the city of Prague.