- February 2018
- December 2017
- November 2017
- October 2017
- September 2017
- August 2017
- July 2017
- May 2017
- April 2017
- March 2017
- February 2017
- December 2016
- September 2016
- December 2015
- November 2015
- September 2015
- August 2015
- July 2015
- May 2015
- April 2015
- March 2015
- November 2014
- October 2014
- September 2014
- June 2014
- January 2014
- December 2013
- October 2013
- August 2013
- July 2013
- May 2013
- April 2013
- March 2013
- February 2013
- January 2013
- December 2012
- November 2012
- October 2012
- August 2012
- July 2012
- June 2012
- May 2012
- April 2012
- March 2012
- February 2012
- January 2012
- December 2011
- November 2011
- October 2011
- September 2011
- August 2011
- July 2011
I have known of RDMA (Remote Direct Memory Access) for quite some time, but never in depth. But since my contract work ended last week, and I have some time off to do some personal development, I decided to look deeper into RDMA. Why RDMA?
In the past 1 year or so, RDMA has been appearing in my radar very frequently, and rightly so. The speedy development and adoption of NVMe (Non-Volatile Memory Express) have pushed All Flash Arrays into the next level. This pushes the I/O and the throughput performance bottlenecks away from the NVMe storage medium into the legacy world of SCSI.
Most network storage interfaces and protocols like SAS, SATA, iSCSI, Fibre Channel today still carry SCSI loads and would have to translate between NVMe and SCSI. NVMe-to-SCSI bridges have to be present to facilitate the translation.
In the slide below, shared at the Flash Memory Summit, there were numerous red boxes which laid out the SCSI connections and interfaces where SCSI-to-NVMe translation (and vice versa) would be required.
NVMe (Non-Volatile Memory Express) is upon us. And in the next 2-3 years, we will see a slew of new storage solutions and technology based on NVMe.
Just a few days ago, The Register released an article “Seventeen hopefuls fight for the NVMe Fabric array crown“, and it was timely. I, for one, cannot be more excited about the development and advancement of NVMe and the upcoming NVMeF (NVMe over Fabrics).
This is it. This is the one that will end the wars of DAS, NAS and SAN and unite the warring factions between server-based SAN (the sexy name differentiating old DAS and new DAS) and the networked storage of SAN and NAS. There will be PEACE.
Nutanix popularized the “No SAN” movement which later led to VMware VSAN and other server-based SAN solutions, hyperconverged techs such as PernixData (acquired by Nutanix), DataCore, EMC ScaleIO and also operated in hyperscalers – the likes of Facebook and Google. The hyperconverged solutions and the server-based SAN lines blurred of storage but still, they are not the usual networked storage architectures of SAN and NAS. I blogged about this, mentioning about how the pendulum has swung back to favour DAS, or to put it more appropriately, server-based SAN. There was always a “Great Divide” between the 2 modes of storage architectures. Continue reading
It has been quite an interesting 2 decades.
In the beginning (starting in the early to mid-90s), SAN (Storage Area Network) was the dominant architecture. DAS (Direct Attached Storage) was on the wane as the channel-like throughput of Fibre Channel protocol coupled by the million-device addressing of FC obliterated parallel SCSI, which was only able to handle 16 devices and throughput up to 80 (later on 160 and 320) MB/sec.
NAS, defined by CIFS/SMB and NFS protocols – was happily chugging along the 100 Mbit/sec network, and occasionally getting sucked into the arguments about why SAN was better than NAS. I was already heavily dipped into NFS, because I was pretty much a SunOS/Solaris bigot back then.
When I joined NetApp in Malaysia in 2000, that NAS-SAN wars were going on, waiting for me. NetApp (or Network Appliance as it was known then) was trying to grow beyond its dot-com roots, into the enterprise space and guys like EMC and HDS were frequently trying to put NetApp down.
“It’s a toy…” was the most common jibe I got in regular engagements until EMC suddenly decided to attack Network Appliance directly with their EMC CLARiiON IP4700. EMC guys would fondly remember this as the “NetApp killer“. Continue reading
Ah, my first blog after Storage Field Day 6!
It was a fantastic week and I only got to fathom the sensations and effects of the trip after my return from San Jose, California last week. Many thanks to Stephen Foskett (@sfoskett), Tom Hollingsworth (@networkingnerd) and Claire Chaplais (@cchaplais) of Gestalt IT for inviting me over for that wonderful trip 2 weeks’ ago. Tegile was one of the companies I had the privilege to visit and savour.
In a world of utterly confusing messaging about Flash Storage, I was eager to find out what makes Tegile tick at the Storage Field Day session. Yes, I loved Tegile and the campus visit was very nice. I was also very impressed that they have more than 700 customers and over a thousand systems shipped, all within 2 years since they came out of stealth in 2012. However, I was more interested in the essence of Tegile and what makes them stand out.
I have been a long time admirer of ZFS (Zettabyte File System). I have been a practitioner myself and I also studied the file system architecture and data structure some years back, when NetApp and Sun were involved in a lawsuit. A lot of have changed since then and I am very pleased to see Tegile doing great things with ZFS.
Tegile’s architecture is called IntelliFlash. Here’s a look at the overview of the IntelliFlash architecture:
So, what stands out for Tegile? I deduce that there are 3 important technology components that defines Tegile IntelliFlash ™ Operating System.
- MASS (Metadata Accelerator Storage System)
- Media Management
- Inline Compression and Inline Deduplication
What is MASS? Tegile has patented MASS as an architecture that allows optimized data path to the file system metadata.
Often a typical file system metadata are stored together with the data. This results in a less optimized data access because both the data and metadata are given the same priority. However, Tegile’s MASS writes and stores the filesystem metadata in very high speed, low latency DRAM and Flash SSD. The filesystem metadata probably includes some very fine grained and intimate details about the mapping of blocks and pages to the respective capacity Flash SSDs and the mechanical HDDs. (Note: I made an educated guess here and I would be happy if someone corrected me)
Going a bit deeper, the DRAM in the Tegile hybrid storage array is used as a L1 Read Cache, while Flash SSDs are used as a L2 Read and Write Cache. Tegile takes further consideration that the Flash SSDs used for this caching purpose are different from the denser and higher capacity Flash SSDs used for storing data. These Flash SSDs for caching are obviously the faster, lower latency type of eMLCs and in the future, might be replaced by PCIe Flash optimized by NVMe.
This approach gives absolute priority, and near-instant access to the filesystem’s metadata, making the Tegile data access incredibly fast and efficient.
Tegile’s Media Management capabilities excite me. This is because it treats every single Flash SSD in the storage array with very precise organization of 3 types of data patterns.
- Write caching, which is high I/O is focused on a small segment of the drive
- Metadata caching, which has both Read and Write I/O is targeted to a slight larger segment of the drive
- Data is laid out on the rest of the capacity of the drive
Drilling deeper, the write caching (in item 1 above) high I/O writes are targeted at the drive segment’s range which is over-provisioned for greater efficiency and care. At the same time, the garbage collection(GC) of this segment is handled by the respective drive’s controller. This is important because the controller will be performing the GC function without inducing unnecessary latency to the storage array processing cycles, giving further boost to Tegile’s already awesome prowess.
In addition to that, IntelliFlash ™ aligns every block and every page exactly to each segment and each page boundary of the drives. This reduces block and page segmentation, and thereby reduces issues with file locality and free blocks locality. It also automatically adjust its block and page alignments to different drive types and models. Therefore, I believe, it would know how to align itself to a 512-bytes or a 520-bytes sector drives.
The Media Management function also has advanced cell care. The wear-leveling takes on a newer level of advancement where how the efficient organization of blocks and pages to the drives reduces additional and often unnecessary erase and rewrites. Furthermore, the use of Inline Compression and Inline Deduplication also reduces the number of writes to drives media, increasing their longevity.
Compression and deduplication are 2 very important technology features in almost all flash arrays. Likewise, these 2 technologies are crucial in the performance of Tegile storage systems. They are both inline i.e – Inline Compression and Inline Deduplication, and therefore both are boosted by the multi-core CPUs as well as the fast DRAM memory.
I don’t have the secret sauce formula of how Tegile designed their inline compression and deduplication. But there’s a very good article of how Tegile viewed their method of data reduction for compression and deduplication. Check out their blog here.
The metadata of data access of each and every customer is probably feeding into their Intellicare, a cloud-based customer care program. Intellicare is another a strong differentiator in Tegile’s offering.
Oh, did I mentioned they are unified storage as well with both SAN and NAS, including SMB 3.0 support?
I left Tegile that afternoon on November 5th feeling happy. I was pleased to catch up with Narayan Venkat, my old friend from NetApp, who is now their Chief Marketing Officer. I was equally pleased to see Tegile advancing ZFS further than the others I have known. With so much technological advancement and more coming, the world is their oyster.
The Riverbed SteelFusion (aka Granite) impressed me the moment it was introduced to me 2 years ago. I remembered that genius light bulb moment well, in December 2012 to be exact, and it had left its mark on me. Like I said last week in my previous blog, the SteelFusion technology is unique in the industry so far and has differentiated itself from its WAN optimization competitors.
To further understand the ability of Riverbed SteelFusion, a deeper inspection of the technology is essential. I am fortunate to be given the opportunity to learn more about SteelFusion’s technology and here I am, sharing what I have learned.
What does the technology of SteelFusion do?
Riverbed SteelFusion takes SAN volumes from supported storage vendors in the central datacenter and projects the storage volumes (aka LUNs)to applications and hosts at the remote branches. The technology requires a paired relationship between SteelFusion Core (in the centralized datacenter) and SteelFusion Edge (at the branch). Both SteelFusion Core and Edge are fronted respectively by the Riverbed SteelHead WAN optimization device, to deliver the performance required.
The diagram below gives an overview of how the entire SteelFusion network architecture is like:
The word “CONVERGENCE” is boiling over as the IT industry goes gaga over darlings like Simplivity and Nutanix, and the hyper-convergence market. Yet, if we take a step back and remove our emotional attachment from the frenzy, we realize that the application and implementation of hyper-convergence technologies forgot one crucial element – The other people and the other offices!
ROBOs (remote offices branch offices) are part of the organization, and often they are given the shorter end of the straw. ROBOs are like the family’s black sheeps. You know they are there but there is little mention of them most of the time.
Of course, through the decades, there are efforts to consolidate the organization’s circle to include ROBOs but somehow, technology was lacking. FTP used to be a popular but crude technology that binds the branch offices and the headquarter’s operations and data services. FTP is still used today, in countries where network bandwidth costs a premium. Data cloud services are beginning to appear of part of the organization’s outreaching strategy to include ROBOs but the fear of security weaknesses, data breaches and misuses is always there. Often, concerns of the weaknesses of the cloud overcome whatever bold strategies concocted and designed.
For those organizations in between, WAN acceleration/optimization techonolgy is another option. Companies like Riverbed, Silverpeak, F5 and Ipanema have addressed the ROBOs data strategy market well several years ago, but the demand for greater data consolidation and centralization, tighter and more effective data management and data control to meet the data compliance and data governance requirements, has grown much more sophisticated and advanced. Continue reading
It’s been a while since I wrote. I had just finished a 2-week stint in Melbourne, conducting 2 Data ONTAP classes and had a blast.
But after almost 3 1/2 months of doing little except teaching NetApp classes, the stint is ending. I wanted it that way, to take a break and also to take on a new challenge. I will be taking on a job with Hitachi Data Systems, going back to the industry that I have termed the “Wild, wild west”. After a 4 1/2-year hiatus, I think that industry still behaves the way it is .. brash, exclusive, rich! The oligarchy of the oilmen are still laughing their way to the banks. And it will be my job to sell storage (and cloud) solutions to them.
In my Netapp (and EMC) engagements in the past 6 months, I have seen the greater adoption of iSCSI over Fibre Channel, and many has predicted that 10Gigabit Ethernet will be the infliction point where iSCSI can finally stand shoulder-to-shoulder with Fibre Channel. After all, 10 Gigabit/sec is definitely faster than 8 Gigabit/sec Fibre Channel, right? WRONG! (I am perfectly aware there is a 16 Gigabit/sec Fibre Channel, but can’t you see I am trying to start an argument here?)
Delivering SCSI data load over iSCSI on 10 Gigabit/sec Ethernet does not necessarily mean that it would be faster than delivering the same payload over 8 Gigabit/sec Fibre Channel. This statement can be viewed in many different ways and hence the favourite IT reply would be … “It depends“.
I would leave this performance argument for another day but today we are going to talk about some of the key additions to supercharge 10 Gigabit Ethernet for data delivery in storage networking capacity. In addition, 10 Gigabit Ethernet is the primary transport for Fibre Channel over Ethernet (FCoE) and it is absolutely critical that 10 Gigabit Ethernet must be close to as reliable as Fibre Channel for data delivery in a storage network.
Ethernet is a non-deterministic protocol, and therefore, its delivery result is dependent on many factors. Likewise 10 Gigabit Ethernet has inherited part of that feature. The delivery of data over Ethernet can be lossy, i.e. packets can get lost and the upper layer application protocols will have to respond to detecte the dropped packets and to ensure lost packets are redelivered to complete the consignment. But delivering data in a storage network cannot be lossy and in most cases of SANs, the requirement is to have the data arrive in the sequence they were delivered. The SAN fabric (especially with the common services of Layer 3 of the FC protocol stack) and the deterministic nature of Fibre Channel protocol were the reasons many has relied on Fibre Channel SAN technology for more than a decade. How can 10 Gigabit Ethernet respond?
It’s been a busy, busy month for me.
And when the IDC Worldwide Quarterly Disk Storage Systems Tracker for 3Q12 came out last week, I was reading in awe how impressive EMC was at the figures that came out. But most impressive of all is how the storage market continue to grow despite very challenging and uncertain business conditions. With the Eurozone crisis, China experiencing lower economic growth numbers and the uncertainty in the US economic sectors, it is unbelievable that the storage market grew 24.4% y-o-y. And for the first time, 7,104PB was shipped! Yes folks, more than 7 exabytes was shipped during that period!
In the Top 5 external disk storage market based on revenue, only EMC and HDS recorded respectable growth, recording 8.7% and 13.8% respectively. NetApp, my “little engine that could” seems to be running out of steam, earning only 0.9% growth. The rest of the field, IBM and HP, recorded negative growth. Here’s a look at the Top 5 and the rest of the pack:
HP -11% decline is shocking to me, and given the woes after woes that HP has been experiencing, HP has not seen the bottom yet. Let’s hope that the new slew of HP storage products and technologies announced at HP Discover 2012 will lift them up. It also looked like a total rebranding of the HP storage products as well, with a big play on the word “Store”. They have names like StoreOnce, StoreServ, StoreAll, StoreVirtual, StoreEasy and perhaps more coming.
The Open SAN market, which includes iSCSI has EMC again at Number 1, with 29.8%, followed by IBM (14%), HDS (12.2%) and HP (11.8%). When combined with NAS numbers, the NAS + Open SAN market, EMC has 33.5% while NetApp is 13.7%.
Of course, it is just not about external storage because the direct-attached storage numbers count too. With that, the server vendors of IBM, HP and Dell are still placed behind EMC. Here’s a look at that table from IDC:
There’s a highlight of Dell in the table above. Dell actually grew by 4.0% compared to decline in HP and IBM, gaining 0.1%. However, their numbers seem too tepid and led to the exit of Darren Thomas, Dell’s storage group head honco. News of Darren’s exit was on TheRegister.
I also want to note that NAS growth numbers actually outpaced Open SAN numbers including iSCSI.
This leads me to say that there is a dire need for NAS technical and technology expertise in the local storage market. As the adoption of NFSv4 under way and SMB 2.0 and 3.0 coming into the picture, I urge all storage networking professionals who are more pro-SAN to step out of their comfort zone and look into NAS as well. The world is changing and it is no longer SAN vs NAS anymore. And NFSv4.1 is blurring the lines even more with the concepts of layout.
But back to the subject to storage market, is there no one out there challenging EMC in a big way? NetApp was, some years ago, recorded double digit growth and challenging EMC neck-and-neck, but that mantle seems to be taken over by HDS. But both are long way to go to get close to EMC.
Kudos to the EMC team for damn good execution!
I was very confused with an recent HDS announcement, and it has been at the back of my mind for several weeks now.
On the last week of September 2012, HDS announced their Hitachi Unified Storage VM, aimed at small/medium enterprises (SMEs). Nothing wrong with that, except the VM part. I am not sure if it was the Computerworld author’s mistake, but he specifically mentioned VM as “virtual machine”. Check out the link here and the screenshot below:
It got me a bit riled up thinking this was some kind of virtual storage ala VMware Virtual Storage Appliance or NetApp ONTAP-V or even the early innovation of HP Lefthand Virtual SAN Appliance. Apparently not!
I did some short investigation and found Nigel Poulton’s blog which gave a fantastic dissection about the HUS VM. The VM is not virtual machine, but Virtual Midrange!
The HUS VM architecture is deep in ASICs, given HDS long history in ASICs design and manufacturing. SiliconFS, is the NAS front end, while the iSCSI and FC part are being serviced from the same HDS microcode of the higher end HDS VSP. Here’s a look at the hardware architectural diagram from Nigel’s blog:
There are plenty of bells and whistles in the HUS VM, armed with plenty of 8Gbps FC ports, SAS 6Gbps backend, SSDs, and software such as Dynamic Provisioning (thin provisioning) and Dynamic Tiering.