This post has already been read 12036 times!
Tech Field Day events bring together innovative IT product vendors and independent thought leaders to share information and opinions in a presentation and discussion format. Independent bloggers, freelance writers, and podcasters have a public presence that has immense influence on the ways that products and companies are perceived and by the general public. November 8th Storage field day has finished, below is my experience:
08:00-10:00 Nimble Storage HQ
The CASL architecture is designed from the ground up to leverage flash and disk together.
CASL’s redirect-on-write snapshots support frequent recovery points with reduced space and performance requirements. The solution’s Intelligent Data Management technology powers WAN replication with configurable retention schedules, and is VM and application consistent.
Nimble Storage till this date has around 1500 total deployments worldwide. Nimble Storage uses Scale-Out Clustering in 2 ways, add storage or add cache / compute power.
Nimble mentioned only needing 12 disks to run 20,000 exchange mailboxes.
Nimble Storage Architecture The hardware differences: The fundamental difference is the storage media. Three Storage tiers: disk only / flash only / disk + flash.
Nimble has three different architectures:
- Disk: Suitable for sequential Application
- Flash Only: Guaranteed high performance, but expensive.
- Flash + Disk: Flash Tier for hot data, disk tier for cold data. most data is cold (old e-mail or snapshots)
CASL writes sequential Stripes to SSD & SATA based on data characteristics and ensures no white space. SSDs are 100% cache for Reads.
A copy of all active or “hot” data (and even “semi-active” data) is held in the large flash layer,enabling very fast random read performance. CASL’s index tracks hot data blocks in real time, adapting to hot spot and application workload changes in a real time fashion. The flash works a JBOD since all the writes are committed.
Performance Policy’s are used to change block size and dedupe / cache size on a per lun basis which is very powerful.
We are at the Sheraton with NimbusData CEO/Founder Tom Isakovich. NimbusData provides an intelligent and efficient 100% flash based storage platform that enables company’s to solve their storage performance problems and reduces the IOP / GB per tile ratio drastically.
Nimbus released their new flagship the Gemini Flash Memory Array. Performance goes up from the E-Class’s maximum of 800,000 IOPS to a million IOPS per Gemini, that’s a cool 20 million IOPS and 1PB all flash storage for a standard rack of these boxes setting a record performance.
2TB in 2,5” using MLC, 24x EFD per Gemini system = 48TB of Flash in 2U’s – w/Redundancy & 10 Year Endurance (8x more than 15k arrays. <10W per TB. 80% less than 15k arrays.
Halo Operating System (datasheet here) Provides unified iSCSI, Fibre Channel, Infiniband, NFS, and CIFS support for block and file storage in one system. Software included is Thin Provisioning, Automatic space reclamation (VMware VAAI) and Inline dedupe. Management via Intuitive GUI, Powerful CLI, and REST API.
Supports NFS 2,3,4; SMB 1,2; SMB 3 in 2013.
13:30-15:30 Zerto (not broadcasted live)
Shannon Snowden – Senior Technical Marketing Architect at Zerto is white-boarding their presentation on Storage Field Day showing their architecture and what their products can do.
Their VRA and their custom hypervisor based replication driver can replicate VM’s from any to any storage vendor and protocol synchronously.
Zerto Virtual Replication (ZVR 2.0) is the first DR solution built specifically to work in private, public and hybrid cloud environments, and is fully integrated with VMware vCloud Director. Check out what’s new here
Tintri marketing VP Geoff Stedman on recent growth: >150 customers, 21K VMs running on Tintri now. Tintri had the following product enhancements for the last 6months: – Per VM snapshot – Per VM Cloning – Certified VMware VAAI Integration
Tintri is VM-Aware storage, The system knows VMware file types so it can really dig deeper into what’s really going on with storage at the VM level. The management software breaks the VM guest down to VMDK and even VM Swap files, giving you granular monitoring and control. The system is purpose build for Virtualized Environments. NFS Only – the product support NFS only.
Performance – You can pin or give more performance to a single VM guest data file if need be.
Tintri is optimized with SSD to reduce latency for reads as well as writes making the none flash drives run as if they were.
Technical deep dive on the Tintri architecture and file system:
Last presenter today is Asigra, Asigra brings a Cloud Model to backup and recovery – including the backup and recovery of mobile devices such as tablets and smart phones.
Version 12 launched yesterday (7 Nov 2012). File based backup, snapshot based backups. I will write more about Asigra as it looks like an awesome solution not many people know of.