ILM – Silverton Consulting

StorageGRID Business is booming

On the call, NetApp announced they sold 50PB of StorageGRID in FY’16 with 20PB of that in the last quarter and also reported 270% Y/Y revenue growth, which means they are starting to gain some traction in the marketplace. Are we seeing an acceleration of object storage adoption?

As you may recall, StorageGRID comes in a software only solution that runs on just about any white box server with DAS or as two hardware appliances: the SG5612 (12 drive); and the SG5660 (60 drive) nodes. You can mix and match any appliance with any white box software only solution, they don’t have to have the same capacity or performance. But all nodes need network and controller/admin node(s) access.

StorageGRID past

Somewhere during Bycast’s journey they developed support for tape archives and information lifecycle management (ILM) for objects. The previous generation, StorageGrid 10.2 had a number of features, including:

S3 cloud archive support that allowed objects to be migrated to AWS S3 as they were no longer actively accessed

NAS bridge support that allowed CIFS/SMB or NFS access to StorageGRID objects, which could also be read as S3 objects for easier migration to/from object storage;

Hierarchical erasure coding option that was optimized for efficiently storing large objects;

Node level erasure coding support that can be used to rebuild data for node drive failures, without having to go outside the node data retrieval;

Object byte-granular range read support that allowed users to read an object at any byte offset without requiring rebuild;

Support for OpenStack Swift API that made StorageGRID objects natively available to any OpenStack service; and

Software support for running as Docker containers or as a VM under VMware ESX, or OpenStack KVM that allowed StorageGRID software to run just about anywhere.

StorageGRID present and future

But customers complained StorageGRID was too complex to install and update which required too much hand holding by NetApp professional services. StorageGRID Webscale 10.3 was targeted to address these deficiencies. Some of the features in StorageGrid 10.3, include:

Radically simplified, more modern UI, new dashboard and policy wizard/editor, so that it’s a lot easier to manage the StorageGRID. All features of the UI are also available via RESTfull API access and the UI is the same for white box, software only implementations as well as appliance configurations.

Simplified automated installation scripts, so that installations that used to take multiple steps, separate software installs and required professional services support, now use a full-solution software stack install, take only minutes and can be done by the customers alone;

S3 object versioning support, so that objects can have multiple versions, limited via the UI, if needed, but provide a snapshot-like capability for S3 data that protects against object accidental deletion.

ILM policy change predictions/modeling, so that admins can now see how changes to ILM policies will impact StorageGRID.

Even more flexibility in DAS storage, so that future StorageGRID configurations can support 10TB drives and 6TB FIPS-140 drive encryption support, which adds to the current drive capacity and data security options already available in StorageGRID.

To top it all off, StorageGRID 10.3 improves performance for both small (30KB) and large (300MB) object get/puts.

Small S3 Load Data Router (LDR, 1-thread) object performance has improved ~4X for both PUTs and GETs; and

Large S3 LDR (1-thread) object performance has improved ~2X for PUTs and ~4X for GETs.

Object storage market heating up

Apparently, service providers are adopting object storage to provide competition to AWS, Azure and Google cloud storage for backup and storage archives as well as for DR as a service. Also, many media and other customers managing massive data repositories are turning to object storage to support their multi-site, very large file libraries. And as more solution vendors support S3 object protocols for data access and archive, something like StorageGRID can become their onsite-offsite storage alternative.

And Amazon, Azure and Google are starting to realize that most enterprise customers are not going to leap to the cloud for everything they do. So, some sort of hybrid solution is needed for the long term. Having an on premises and off premises object storage solution that can also archive/migrate data to the cloud is a great hybrid alternative that takes enterprises one step closer to the cloud.

Comments?

Susitna Glacier, Alaska by NASA Goddard Photo and Video (cc) (from Flickr)

Talk about big data, Technology Review reported this week that IBM is building a 120PB storage system for some unnamed customer. Details are sketchy and I cannot seem to find any announcement of this on IBM.com.

Hardware

It appears that the system uses 200K disk drives to support the 120PB of storage. The disk drives are packed in a new wider rack and are water cooled. According to the news report the new wider drive trays hold more drives than current drive trays available on the market.

For instance, HP has a hot pluggable, 100 SFF (small form factor 2.5″) disk enclosure that sits in 3U of standard rack space. 200K SFF disks would take up about 154 full racks, not counting the interconnect switching that would be required. Unclear whether water cooling would increase the density much but I suppose a wider tray with special cooling might get you more drives per floor tile.

There was no mention of interconnect, but today’s drives use either SAS or SATA. SAS interconnects for 200K drives would require many separate SAS busses. With an SAS expander addressing 255 drives or other expanders, one would need at least 4 SAS busses but this would have ~64K drives per bus and would not perform well. Something more like 64-128 drives per bus would have much better performer and each drive would need dual pathing, and if we use 100 drives per SAS string, that’s 2000 SAS drive strings or at least 4000 SAS busses (dual port access to the drives).

The report mentioned GPFS as the underlying software which supports three cluster types today:

Shared storage cluster – where GPFS front end nodes access shared storage across the backend. This is generally SAN storage system(s). But the requirements for high density, it doesn’t seem likely that the 120PB storage system uses SAN storage in the backend.
Networked based cluster – here the GPFS front end nodes talk over a LAN to a cluster of NSD (network storage director?) servers which can have access to all or some of the storage. My guess is this is what will be used in the 120PB storage system
Shared Network based clusters – this looks just like a bunch of NSD servers but provides access across multiple NSD clusters.

Given the above, with ~100 drives per NSD server means another 1U extra per 100 drives or (given HP drive density) 4U per 100 drives for 1000 drives and 10 IO servers per 40U rack, (not counting switching). At this density it takes ~200 racks for 120PB of raw storage and NSD nodes or 2000 NSD nodes.

Unclear how many GPFS front end nodes would be needed on top of this but even if it were 1 GPFS frontend node for every 5 NSD nodes, we are talking another 400 GPFS frontend nodes and at 1U per server, another 10 racks or so (not counting switching).

If my calculations are correct we are talking over 210 racks with switching thrown in to support the storage. According to IBM’s discussion on the Storage challenges for petascale systems, it probably provides ~6TB/sec of data transfer which should be easy with 200K disks but may require even more SAS busses (maybe ~10K vs. the 2K discussed above).

Software

IBM GPFS is used behind the scenes in IBM’s commercial SONAS storage system but has been around as a cluster file system designed for HPC environments for over 15 years or more now.

Given this many disk drives something needs to be done about protecting against drive failure. IBM has been talking about declustered RAID algorithms for their next generation HPC storage system which spreads the parity across more disks and as such, speeds up rebuild time at the cost of reducing effective capacity. There was no mention of effective capacity in the report but this would be a reasonable tradeoff. A 200K drive storage system should have a drive failure every 10 hours, on average (assuming a 2 million hour MTBF). Let’s hope they get drive rebuild time down much below that.

The system is expected to hold around a trillion files. Not sure but even at 1024 bytes of metadata per file, this number of files would chew up ~1PB of metadata storage space.

GPFS provides ILM (information life cycle management, or data placement based on information attributes) using automated policies and supports external storage pools outside the GPFS cluster storage. ILM within the GPFS cluster supports file placement across different tiers of storage.

All the discussion up to now revolved around homogeneous backend storage but it’s quite possible that multiple storage tiers could also be used. For example, a high density but slower storage tier could be combined with a low density but faster storage tier to provide a more cost effective storage system. Although, it’s unclear whether the application (real world modeling) could readily utilize this sort of storage architecture nor whether they would care about system cost.

Nonetheless, presumably an external storage pool would be a useful adjunct to any 120PB storage system for HPC applications.

Can it be done?

Let’s see, 400 GPFS nodes, 2000 NSD nodes, and 200K drives. Seems like the hardware would be readily doable (not sure why they needed watercooling but hopefully they obtained better drive density that way).

Luckily GPFS supports Infiniband which can support 10,000 nodes within a single subnet. Thus an Infiniband interconnect between the GPFS and NSD nodes could easily support a 2400 node cluster.

The only real question is can a GPFS software system handle 2000 NSD nodes and 400 GPFS nodes with trillions of files over 120PB of raw storage.

As a comparison here are some recent examples of scale out NAS systems:

EMC Isilon demonstrated a 140 node storage cluster in a SPECsfs2008 benchmark.
IBM has already announced support for a 14.4PB SONAS storage system using high density drives

It would seem that a 20X multiplier times a current Isilon cluster or even a 10X multiple of a currently supported SONAS system would take some software effort to work together, but seems entirely within reason.

On the other hand, Yahoo supports a 4000-node Hadoop cluster and seems to work just fine. So from a feasability perspective, a 2500 node GPFS-NSD node system seems just a walk in the park for Hadoop.

Of course, IBM Almaden is working on project to support Hadoop over GPFS which might not be optimum for real world modeling but would nonetheless support the node count being talked about here.

——

I wish there was some real technical information on the project out on the web but I could not find any. Much of this is informed conjecture based on current GPFS system and storage hardware capabilities. But hopefully, I haven’t traveled to far astray.