The last few days were insanely busy, with sessions and discussion and emerging new standards, but all in all Tech Field Day Seattle was an amazing experience. I took a pile of notes, though there could have been more and met a group of great people who I would likely not have been able to meet otherwise. We heard from several companies and got great information about using new ideas and products and spent a lunch hour discussing the effectiveness of a new standard in networking communication #FCoTR (Fibre Channel over Token Ring).
This post is an overall view of what I got out of #Techfieldday, which was quite a lot.
What’s this all about?
As in a previous post, Tech Field Day is a summit of technical thinkers and organizations who can benefit from the feedback each has to give. I realize that the sponsor companies were looking for product feedback and constructive discussion which they got, but I think the delegates were able to learn a great deal as well, both about the hot new products coming from the sponsor companies and about technology in general.
I know I got a lot out of the networking opportunity with everyone and that was very valuable for me.
Getting into the discussion
We heard from 5 companies at Tech Field Day Seattle this week and the theme was definitely storage. The companies that came out were:
Tech Field Day Seattle Sponsor Companies
Virtualization of both systems and storage were hot topics this week and rightfully so. Since big storage allows a huge amount of virtualization possibilities the two go hand in hand.
Working to provide feedback to these organizations as well as have some open discussion about our opinions of the product was the overall goal of the sessions. To be honest and open and not solely “for the press” by giving the presenters honest feedback and not worrying about media image, but of technical relevancy was a nice change.
A bit about the companies we saw
First up was Veeam. Its always good to go first, sets the tone for the rest of the day.
The Veeam crew provided information about their Virtualization solutions providing backup and replication of virtual machines. The seemingly easy to use process of being able to get your VM Snapshots replicated to a DR environment. Being new to the Virtualization space (and wishing spell check would be OK with the word Virtualization) I can see the usefulness of a tool that takes the guesswork out of ensuring your VM environment is both backed up and safe somewhere else in the environment.
But some would say…
Snapshots are not backup. Which they arent, technically. However if there are enough snapshots taken regularly, there is no reason that this could not be almost a complete replacement for actual backup. Depending on the size of the environment and resources available.
My organization is just now looking into Virtualization and this solution could be a great supplement to existing backup methods once we get there. Perhaps to keep the VMs in a mount-ready state at another site and back them up with our current solution.
Next up was Nimble Storage. They built an array that uses both flash based storage and more affordable storage to provide a high-speed read/write solution and manage it all from one interface and in one box.
This was a highlight for me. Not only because I am interested in and learning about storage arrays but because the company launched at Tech Field Day. They have been working on development of the product and with customers, but chose this event to go live.
Bringing to market a storage solution is one thing, but to deliver a product that can accomodate what appear to me to be the pitfals of typical storage, fast reads and lower cost long term storage could be huge.
Most companies like cheap storage… cheap is relative of course, but being able to put data on SATA disks can save quite a bit of money in overall solution costs. They are cheap and they work great for read operations. So with SATA disk you get great storage and read access, but the writes can be a bit slow. Depending on the environment and the number of drives in the solution (and number of hits by users) the reads may also be slow, but many times not deal-breaking.
Along with the afford-ability of backend storage, Nimble Storage combines high performance writes to the array in the form of a flash cache. On the front of the storage system there is a flash based cache that allows nearly instantaneous access to storage (both for read and write). Data goes into the cache and is written to storage or disk at intervals. So if there is a file that gets written into the cache and is highly used it may not get written until much later in the day making it instantly available within the flash disk for your users.
I think of it as a tiered approach (although it is not quite the same as other tiered solutions we saw, more there in a bit) where information goes into the flash based portion of the solution on write and stays there until the cache gets to a certain percent full, then is written to the disk. Once the cache is written to disk portions of it can be cleared to make room for other data.
Tiered storage isn’t new
I know there are other organizations that use tiering to keep heavily accessed stuff highly available to the users, which is a great concept, but the heavy incorporation of a set of SSD media to improve overall performance (especially writes to the array) is great.
After lunch on day one, we took a break from the Storage/Backup stuff for a while and went over to F5 networks for a session.
F5 was new to me. I had seen their building before on other visits to Seattle, but wasn’t really sure what they were all about. What I learned was networks were the business they were in but it wasnt necessarily pointing users or companies at the Internet (or network), it was more about providing an experience on the network for the users that was optimized in many ways.
The F5 core product is a device called Big IP which proxies connections to resources and can optimize the availability or access to these resources for the user needing them. This was very cool.
An example would be a website that lives on servers spread across multiple locations. The Big IP can optimize the use of these resources by doing things like figuring out which web server is closest to the requester and routing the request to that server to display the content. Doing this creates better resource distribution for the servers and a much faster/smoother experience for the user.
Another concept that was introduced to me while at F5 (aside from the awesome cupcakes) was their technologies of iRules and iControl and the community they have put together to help provide and support these technologies.
iRules are the items that move things around. They allow you to apply scripts to traffic that hits the Big IP device and take action on it. Like the delivery of content from a server near you as discussed in the above example or working around an error found on a web page. Suppose my website was sitting behind a Big IP device and I discovered that a string entered in browser when looking for certain content allowed pages to be manipulated by users. iRules could be created to inspect the traffic as it was coming through the Big IP and based on the rule send the visitor to a predetermined place instead of the area where the manipulation might happen. This effectively works around the issue.
iControl allows applications on a network to work together even on dissimilar environments. Traffic monitoring and business process management come to mind for me here. Suppose you wanted to monitor traffic coming into your network across all locations. iControl would be able to provide a simple interface across all of these devices to allow for monitoring rather than having to capture all the data at each location and work with it separately.
Tech Field Day – Day 2
On day two we were able to board the bus a bit later, which was great. First on day 2 was Compellent for a session about their storage technology.
Compellent produces storage arrays that support automatic tiering of data across multiple types of disks. This is really pretty sweet. The automatic tiering idea is one that takes a look at the data being stored and scrubs it for statistics every day. With the information it discovers during scrubbing the algorithm used within the array can determine how much use their is on a file (or group of files) and move limited use data to a higher tier of storage, or one that isnt as frequently used.
Perhaps your production database is kept in tier 1 to allow for fastest access, but some of your company’s marketing material is not accessed all that much and can be kept on tier 3 storage costing less money long term.
Why do I care about where items live?
If an item is used infrequently it can live on inexpensive (and slower) storage media. Perhaps things that are not used often will live on SATA drives within the array and things used every day will live on faster SAS drives. This could be done by administrators, moving things around to optimize their use, but Compellent takes care of this and does it all within the same array.
Another area where I found Compellent to be particularly impressive was in replication. With Compellent arrays on both ends of a DR solution, replication between these arrays can be done at a much faster rate because the information can be compressed and only the changes replicated rather than the entire file during certain time periods. This can speed up performance and optimize bandwidth when needed by users.
During the intermission…
During the intermission of lunch and blogging at TFD day 2 a great deal of discussion was had covering a new networking standard Fibre Channel over Token Ring. You can read more about that here. We also made a trip to building 92 at Microsoft to check out the company store and more of their campus.
Back to Storage
Our last presentation was at NEC North America and covered their HYDRAStor product.
When I got to NEC I had no idea they were in the storage business. I have used NEC projectors and things but this was completely surprising to me. Not only was the product new to me (and most apparently) but our host Gideon Senderov was a wealth of knowledge on the product. It was by far the presentation that captured my attention the most.
The HYDRAStore product is built on NEC gear (Servers… who knew NEC made servers). And each device making up the solution, the Accellerator Nodes and the Storage Nodes are running on servers with memory to facilitate access and improve performance.
Midway through the presentation we toured the test lab at NEC which seemed to get its cooling right out of a wind tunnel. No the HYDRAStor arrays do not need a wind tunnel for cooling, but it was kinda cool to see it in place. The lab had four racks full of storage running and was able to hit throughput rates at 10 GB per second. The speed offered here was definitely amazing.
The NEC Solution was very high end and seemed to me to be greatly suited for archival. Something used to keep long term copies of your data on with a suplimental storage solution used for every day read/write.
Overall thoughts on Tech Field Day
I had a blast networking with the other delegates and sponsor companies at Tech Field Day Seattle and learned quite alot about these companies and their products. Coming from an end user role this helped me to see why Storage is so paramount and that there are many ways to skin the storage cat.
I hope the sponsors got the feedback they were looking for and will continue to support Tech Field Day in the future. As for the solutions, they are all extremely good and accomplish big things. I think combining some of them, perhaps in a very forward thinking environment or a large environment could be a great all around storage and or backup solution.
I am working further with information learned at Tech Field Day and will be communicating with the sponsor companies to further develop my notes for future blog posts.
Finally I would like to thank Stephen Foskett and others at Gestalt IT for allowing me to be a part of Tech Field Day and the other delegates for helping me learn so much while attending.
use backing up data is something that goes without question. Sure care should be taken to select the right software and media to ensure your (or your organization’s) data is regularly backed up. Imagine if there was a single solution allowing easy management and usage.
