12_tribes - Fotolia
Storage Decisions 2015 in New York last week ended with a panel of experts who took attendees questions about storage technologies, strategies and vendors, and how it's affecting today's data storage industry.
The Storage Decisions experts included Jason Buffington of Enterprise Strategy Group, George Crump of Storage Switzerland, Marc Staimer of Dragon Slayer Consulting, Jon Toigo of Toigo Partners International, and Ben Woo of Neuralytix. The questions involved cloud storage, flash, spinning disk and even tape.
Here are the highlights from that question and answer session:
What are the advantages and disadvantages of cloud storage?
"Flexibility, convenience and, in some cases, cost are the main drivers of the cloud," Woo said. "It's a great vehicle for disaster recovery, business continuance."
"The real advantage of cloud storage is not cost. It will cost you more," Staimer said. "The real advantage is flexibility. When you have data in the cloud, you don't have to have your own infrastructure. You don't have to have people managing infrastructure for this huge amount of data, specifically archive data. The downside is cost. Once it's there, it's sticky. Moving it off the cloud is a time-consuming process, and once you move it into the cloud, you're probably never going to move it out."
Toigo pointed out that the "Cloud is dependent on the network and bandwidth that connects you to the service. When you need your data, the network better be there to allow you to move it in an efficient manner. Migrating from one cloud to another cloud is a costly endeavor, more costly than migrating from one generation of tape to another generation of tape."
He also predicted low costs of the public cloud due to price wars between Amazon, Microsoft and Google won't last.
"The cost of cloud storage is as cheap as it is because of the price war," he said. "It's not sustainable. The price will adjust and everything will get more expensive."
With all the changes going on in the data storage industry and with different technologies and vendors, we need to make decisions now. What should we buy now?
"Technology's always going to be changing," Staimer said. "You have to ask yourself, 'What are my requirements, now, next year, the year after, and the year after that, and how do I plan for it?' You can only plan with the knowledge you have, so you have to recognize whatever you plan for now, you may be changing that as time goes on. It depends on your requirements."
"We can't do an audit of your company now, but I think we can make some generalizations," Crump said. "For primary storage, you probably want to have some flash. You might want to have all-flash. If you have archiving, that's environmental, and then from there some sort of tiering strategy ... I believe in tape and you might want cloud."
Is spinning disk dead?
"No. It will be here for a long time," Woo said.
"There's huge install base for spinning disk, and a huge manufacturing capacity for spinning disk," Toigo said. "Seagate is making enormous investments in spinning disk to keep it alive. It's currently the main storage for a preponderance of desktop users and a preponderance of small offices and home offices. There's always going to be a thriving disk market. Is it going to be the primary storage medium for fastest cache? No, that's probably not going to be the case. However, there is a place in the ecosystem for every piece of storage media. Are drives going to maintain a rate where we double capacity every year at the same cost? No, that probably won't happen in the foreseeable future. But I think we'll still have a robust disk market and anybody telling you anything else is pulling your leg and trying to sell you flash."
"OK, I'm pulling your leg," said Staimer, who said he can foresee the demise of disk. "I do see an end of life for disk drives as a major storage medium. One of the fallacious arguments is there is not enough flash today to produce the capacity you need. That's fallacious because it's based on demand.
"Look at it from a capacity and density perspective. By the end of the decade, the difference of what you can get in a flash drive and hard disk drive will be enormous. By 2020, the top capacity of 3.5-inch form factor hard disk drives will be about 20 TB. That's best case scenario. Worst case on the flash side is you'll have a 2.5-inch form factor that's 148 TB."
How can we keep up with demands to keep data for years or even forever?
Buffington said a Storage Decisions attendee told him, "One guy from legal came up to me and said 'If you ever destroy another tape again, you're fired.'"
"That's a knee jerk reaction from a lawyer who is constantly worried he will not find the data he needs," Buffington added. "What they're struggling with is [that] you have to retain the data as long as you need it, but those same lawyers two years from now are going to find data that they wish wasn't there anymore. If data has long-term value you keep it, but the value of that data has to be greater than the liability of disposing it."
"If there's data you don't want, there's a trashcan for that," Toigo said. "Why are we spending a penny a month a gigabyte to store data with a third-party provider that probably can't deliver it back to you when you need it? We're going to have to get serious about this because we're amassing too much data."
Staimer said: "If your manager tells you that you can't throw anything away, you need to establish what that's costing you. And you shouldn't keep data one iota longer than compliance requires to keep it. You need to have rules on how long you keep data and set policies, so it's done. You've not only deleted it, but you securely deleted it."
Crump disagreed with the "delete everything" approach.
"I have no problem with you deleting emails that say 'do you want to go to lunch tomorrow?' That's fine," he said. "But I can't tell you how many times I've seen situations where data was deleted under corporate policy and three years later, it comes back that we could've used that data -- especially if you're talking about big data analytics, and the ability to analyze old sets of data. The ability to wrap this whole data into a big picture in a broader spectrum is invaluable. I totally agree you should delete data, but you have to be very careful about what you're deleting."
What is the best technology for archiving and long-term data preservation?
"To me, it's tape," Toigo said.
"It has to be tape," Crump said. "You must be able to tier data, and tape is the most logical place to put it."
"Tape will be here for a while, at least for decades," Staimer said. "I can see an end of life for hard disk drives. I can see an end of life for flash drives because they will be replaced by other non-volatile memory. But I don't see an end of life for tape."
How will the Dell-EMC deal impact the data storage industry?
"They are the evil machine corporation and they should have died a long time ago," said Toigo, obviously not an EMC fan. "But that's not the point. You were asking about what happens to the storage market when the number one player dissolves. The fact is EMC hasn't been in the storage business for a long time. EMC is 80-something percent owner of VMware, and the money that has been feeding their coffers comes from services around the sales of equipment that they've already sold, and revenues they've received from VMware. [Dell-EMC] may impact the software-defined storage space in a meaningful way, but we have no idea what that meaningful way is going to be."
The other analysts had guesses about that meaningful impact.
"There will be two waves of hesitation in the way this comes together," Buffington said. "There's everything up until the day it closes. And during that time there will be indecision and hesitation by both sets of channels, both sets of customers, so that customers and partners won't know what products to make bets on. Then the second wave of hesitation comes in because of the overlap in technologies. So which technologies are going to lapse, which will come together? How long will it take "DMC" to figure out which products are going to stay? They have to get their story together about two sets of products and two sets of channels, and then they have to recoup customers who jumped ship during that time. They'll have four to six bumpy quarters before things start to settle down."
"I think the real impact will be a lot of smart people from Dell and EMC will leave, start new startups... and we'll see a whole new crop of storage vendors on the market in the next four or five years," Crump said.
"EMC and Dell combined represent 33% of the storage market today," Woo said. "The net result [of the merger] will be loss of market share ... that means more opportunity for emerging vendors."
What about the HP split?
"I'm positive about the split because it does focus HP's businesses in the right places," Woo said. "Sunsetting Helion for instance, will focus them on what they do, and that's infrastructure. Size-wise, HP will be on par with IBM and the Dell-EMC combination. It's going to be an interesting next few years as they compete on much more level playing field and have a more precise focus on what they do."
"The real benefit of the HP split is they no longer have the safety net of the printing division providing them with a cash underpinning," Staimer said. "If they screw up in storage or servers or whatever, they can hide it now. They can't do that anymore [after the split]. Basically they've got to make it now as an enterprise company. They didn't have to do that before. Which is why things like their cloud offering -- Helion -- is gone because it was a non-money maker. I don't know if they're going to succeed but you can expect they will be more customer-oriented and there will be change."
"This is an opportunity for HP to decide what they want to be when they grow up," Buffington said. "This is an opportunity for them to redefine themselves under a new brand."
"The split doesn't change culture," Crump said. "Now you have two companies with two bad cultures."
Predictions for hottest data storage industry technologies of 2015
A four-part strategy for archiving data in place
How the data storage market evolved in 2014