Will Sun continue to resell Hitachi's Lightning products in the high-end space or will Sun be developing its own high-end box? Our three-year relationship has been renewed and we envision our relationship to be long and fruitful. We are selling a big percentage of HDS. One year [into the deal] we made tremendous progress in the high-end space. We could not do that without a product like [HDS']. If you look at what's happening in the...
data center, you have a class of center that an HDS high-end box can easily satisfy. You have centers beyond that, where the high-end [products] need to be aggregated and below that you have the smaller data centers. You really need to focus on how do you virtualize the data centers and provide services. Where does Sun spend its R&D dollars? Our research and development is focused on storage management. Thirty percent of our R&D is going into that space right now. Whatever happened to HighGround? HighGround is very much alive and has helped broaden our software portfolio into Storage ONE. HighGround and the team from HighGround is responsible for delivering the StorEdge Resource Management Suite that allows customers to manage, plan and implement their storage assets, providing higher levels of application services while lowering the cost of storage management. The acquisition has been integrated into our Complete Storage Solutions portfolio. What argument does Sun make to its customers for its new file system and other storage software products versus the Veritas products that Sun has successfully OEM'd for years? The Sun StorEdge QFS software is a high-end file system, opposed to the Veritas File System which is targeted for midrange systems. StorEdge QFS enables true file sharing in a SAN environment along with the highest scalability of up to 252 TB. What is Sun's stance on IP storage? We see a lot of hesitance and a lot of immaturity in the marketplace. I'm looking at the standards constantly. We see the level of stability the [iSCSI] specification reaches and the technology is not there yet. The specification will be ready for deployment in late 2003 or early 2004. The iSCSI hype is fading away. There are a couple of things that can happen to show if there is staying power for the technology or if it will fade away.
There are interesting technologies coming. RDMA over IP for low latency access, which is similar to what InfiniBand offers over a different transport layer. It's IP-based but it is a different technology than iSCSI. We are not a big believer of iSCSI. Sun has a very embryonic effort to keep track of the technology and do some prototyping. How does Sun view its current competition in the storage market?
It's one of those complicated pictures. I think we have the benefit of not being a strong player in the high-end storage market. It's very difficult for EMC to come up with a strategy that doesn't undermine their high-end offerings. It wasn't forethought for us, it was luck. We're in an interesting transition right now. We don't have a lot of legacy hardware to carry forward. I think HP is a more interesting [competitor] because it came from similar line of thinking, but it is yet to be seen how HP is going to execute.
On the EMC side there's an interesting dilemma. They have to compete with the HDS 9980 series platform in the high-end while there is a change percolating underneath them. There are a bunch of startups out there threatening to change the storage paradigm by producing smaller, scalable storage systems. It's always a horse race. The Storage Networking Industry Association recently formed a group dedicated to storage security. What is Sun's stance on the current state of security in storage?
iSCSI security problems are coming from the fact that if you have a single IP network or a unified IP technology for LAN and SAN in a data center. With that you run the risk of XML access allowing break-ins into your storage network. If you look at the larger proportion of the data center, the problem is human error. People are setting things up the wrong way. Fibre Channel has native security because it's not IP-based. VLANs are good for that. That's not to say that Fibre Channel security is efficient. It works OK. You can configure FC devices in-band and out-of-band. If someone breaks into a server they can create access at the kernel level and tunnel their way into the storage and make changes. SAN security definitely needs to be improved. It's very important for the storage networking folks to develop processes and products for security. Does Sun support open standards like CIM, WBEM and Bluefin?
What we wanted to do with JIRO was create the first interoperability standard. It had a lot of features that were still missing. Our push was to define a tech that could be adopted by anybody and provide interoperability in a data center. It's a heterogeneous world and customers want us to make it work. We believe that what's important is to have open APIs and proprietary implementations. Underneath the APIs the implementation should be proprietary. What does the future hold for the storage industry?
Everybody should brace for a major shakeout. There's going to be a large shakeout in the startup space because it's very clear that there are a limited number of channels to sell products to end users. End users aren't going to buy from startups. What's going to happen in the storage space will be more competitive than network space. A lot of the startups are going to be disappointed.
The second shift is around data center virtualization. Storage is going to become more tightly integrated into the data center. Storage consolidation and data center virtualization becoming a single, integrated solution. Let us know what you think about the story, e-mail Kevin Komiega, News Writer
FOR MORE INFORMATION:Sun makes another storage push; targets EMC Sun to roll out SAN tools Users bogged down in IP storage hype Comment on this article in the SearchStorage Discussion forums